December 3, 2024 • 53 mins
Join Àlex Rodríguez Bacardit (CEO @ MarsBased) and John O’Nolan, founder of Ghost, as they delve into the world of independent, values-driven tech businesses. They discuss Ghost’s foundational structure, the role of transparency and bootstrapping in shaping the company, and how these principles impact product development and user loyalty. The conversation also touches on the challenges of competing with heavily funded competitors, the nuances of running a foundation, and the future of newsletter platforms.
Whether you’re an entrepreneur or a tech enthusiast, this episode offers deep insights into aligning business with core values.
Also, hot takes with hot salsas 🔥
🎬 You can watch the video of this episode on the Life on Mars podcast website: https://podcast.marsbased.com/
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Àlex Rodríguez Bacardit (00:00):
Hello everybody.
I'm Alex, founder of MarsBased,and in this episode we bring
you John O'Nolan from Ghost, acompany behind the popular open
source newsletter and bloggingplatform that we have used here
and there at Marsbase.
I have used extensively as mypersonal website over the years.
My founder to foundernewsletter nowadays is built on
Ghost.
Ghost and with John wediscussed many things that we
(00:23):
have got in common as a companythe transparency, the core
values, remote work,bootstrapping, but also some
particularities they've got.
They are a foundation.
It's a kind of company thatcannot be sold.
They will not be able to takethe usual investment that other
companies might be entitled to.
We speak about the future.
(00:44):
We also talk about theirtechnological stack and we dive
a little bit deep into thetechnical intricacies of their
product, particularly regardingrewrites, technical depth and so
on.
And the question of the day forthis episode will be what
newsletter platform do you usefor your company?
John O'Nolan (01:06):
At.
Àlex Rodríguez Bacard (01:07):
MarsBased .
We're considering switchingover from MailChimp to something
else, and your answers will beparticularly helpful, and so,
without further ado, let's jumpright into this episode.
Good, I think this is recording.
You don't look very nervous.
How are you doing?
John O'Nolan (01:23):
John, welcome to the show.
Thank you for having me seemslike you're used to this.
Àlex Rodríguez Bacardit (01:28):
Not my first podcast, but I'm sure it
will be the best you look moreof a professional podcast
recorder than I do, so it seemslike you're very used to this.
I think it's part of the remotesetup, right?
John O'Nolan (01:41):
yeah, that's right .
After a while you get used tosetup right.
Yeah, that's right.
After a while you get used totrying to have a nice setup and
making video calls feel good.
Àlex Rodríguez Bacardit (01:50):
I wish my team could hear this so that
they also bought the nicecameras and mics and they don't
have to be using these setupsthat introduce a lot of echo
into the calls and stuff likethat Good.
I have so many questions.
I've been using Ghost for 10years, on and off.
I have to say that my journeyhas been interesting with the
(02:10):
platform.
I remember trying it too early,but more recently I have
rediscovered it as a personaluser.
I recreated my website Everytwo years.
I've got this thing where Irebuilt my personal website with
a different framework and I waseager to do it with ghost this
time around.
I have to say there's somethingin it that gave me the wow
(02:33):
impression every step of the way.
I think you guys nail it, andso here goes my gratitude and my
you know I'm super impressed byit.
How much of this is affected bythe fact that you're a bootstrap
company, independent company,and that you can take however
long it takes to polish all thedetails.
John O'Nolan (02:54):
Well, I think it is affected by us being a
bootstraps company, but notnecessarily because we can take
all the time to polish thedetails.
You know we have a lot ofcompetitors, pretty big ones I
think the smallest has $50million of funding and the
largest has $1.1 billion offunding and people, as in end
(03:14):
users, don't particularly careif we're bootstrapped and
independent.
They just want to use the bestproduct right.
So if we took all the time inthe world to just polish
everything and make it perfect,we simply wouldn't be able to
keep up with the expectationsthat people have from products
like Ghost.
But that said, I think beingindependent and bootstrapped
allows us to have more focus,allows us to optimize for
(03:36):
different things, allows us totake different strategic moves
than we might otherwise if wehad investors who were pushing
for growth targets Exactly.
Àlex Rodríguez Bacardit (03:47):
I take it that companies like Basecamp,
ghost, buffer you know thatyou're more well known for being
highly opinionated,bootstrapped or
semi-bootstrapped.
You can take more time topolicies, and I know you're
right.
I grant you that you don't haveall the time in the world.
However, you are aware of asubset of users that they will
(04:09):
buy your product, whatever ittakes they don't really care
about.
Maybe there's a bettercompetitor, maybe there's a
cheaper alternative, but I willbuy your product because of your
values, your vision, and I'mtotally aligned with your
product.
For instance, in our companyvalues, your vision and I'm
totally aligned with yourproduct.
Like, for instance, in ourcompany MarsBased.
We created the company becausewe read the books of Basecamp
(04:29):
and so for 10 years we werepaying for that product and we
knew there was better softwareout there.
But we said, like, thanks toRails, thanks to Basecamp, we're
here and now right.
So it took really long for usto kind of like, get rid of
Basecamp when it was, you know,unavoidable that we had to use
other products.
But I find myself usingproducts because I really
(04:50):
believe in the vision, right?
So have you calculated how muchof a percentage of your user
base is this fanatical user basethat will pay for your product.
John O'Nolan (05:02):
you know, whatever you do, or not really Haven't
calculated the percentage.
I think that might be quitedifficult to do.
There's definitely some peoplewho are very passionate about
Ghost based on structure andvalues, and that's a really
important part of how we dothings.
I think the killer combination,though, is if you can combine
the two.
If you can have great structureand values as well as a great
(05:22):
product, I think then you'rereally onto something right.
I think of products outside orcompanies outside of the tech
space with things like thisright.
So, like Patagonia is theclassic example, people are
aligned with their values.
People buy Patagonia stuffbecause of their values, but if
they made crappy t-shirts thatfell apart after a while, then
(05:43):
it wouldn't be what it is today.
It's the quality and the carethat goes into the product,
combined with the values.
That, I think, is where themagic can happen, if you get it
right, exactly, you mentioned it.
Àlex Rodríguez Bacardit (05:55):
You created a brand, and one of the
things about creating a brand ispeople will follow it more
strongly than anything else.
People will be or a user basewill be less volatile, if you
will, but that doesn't happen onday one.
That doesn't happen on day 10.
Are you aware of the day wherethat happened, when you made the
click and you started gettingit right, everything you made,
(06:16):
and you started attracting thissubset of users, the cult-like
users that we were talking aboutbefore?
Are you aware of the precisemoment or movement you did in
the company that kind of, likeyou know, turned the direction
of the company into the rightway?
John O'Nolan (06:32):
Yeah, I think Ghosts is an unusual case where
it did happen on day one, andthat's actually to do with the
origin story of the company,which kind of started as just
this blog post pitch of what Iwanted to exist in the world and
the values that I thought itshould have, and so the values
for the company existed beforethe product, before the company,
(06:54):
they existed as an idea and aconcept that I put out into the
world, and so it was actuallythat that captured people's
attention as much as the ideafor the product itself.
So, as we went from blog poststo Kickstarter to the first
launch, the consistent storythat came along with Ghost was
that we were doing this for love, not for profit, doing this for
(07:15):
values, not for growth, andthat was a huge part in the
launch and then also a huge partin the continued growth that
came afterwards.
Àlex Rodríguez Bacardit (07:24):
Exactly .
One of the things that reallyresonates with me is it's very
similar to Basecams until theend of the internet, right, the
fact that you are a foundation,I think, is something that
stands out.
You are a different kind ofcompany, not only legally, but
also existentially.
If you create a company thatyou cannot sell, you will not do
things that will drive youtowards selling the company,
(07:47):
right, and so you operatedifferently.
Beyond the obvious consequencesof running a foundation, that
might not be obvious for 99% ofthe users of the internet, but
you know, in terms of governance, in terms of leadership.
In terms of leadership, haveyou found, for instance, any
restrictions of things youwanted to do, something that has
(08:08):
held you back because you werea foundation, perhaps, like you
wanted to run faster, which is Idon't believe it's in your
ethos as a company, but whatthings did you find out along
the way that were ah, this isholding me back.
John O'Nolan (08:22):
Yeah, there's a bunch of weird non-profit
regulations and things that youget surprised by, and every
country kind of has their ownset of them, so they'll be
slightly different depending onwhere you are in the world.
The one that comes to mindalways immediately is if you're
incorporated as a non-profitfoundation, you are
(08:42):
automatically classified bybanks and financial institutions
as high risk which is fairlylogical, you know, like most
non-profits are dependent onoutside funding, typically don't
sustain their operationswithout some kind of outside
help.
So do you really want to lendmoney to an organization that
may be not able to pay it back?
As a bank, probably not, butit's a blanket rule.
(09:06):
So, for example, a common thingthat software startups do is
they put their hosting billswhich are usually quite
extensive on American Expresscards, and the benefit of that
is you get a delayed timeframefor paying your hosting bills.
But you also rack up a hugenumber of air miles and points
and things that you can useelsewhere.
Amex won't talk to us.
They say oh, you're anon-profit, no, you're
(09:29):
disqualified.
And same with banks.
We haven't ever needed lendingfacilities, but if we did,
they're not available.
Banks won't give you any linesof credit as a non-profit
organization.
So the flexibility there issomething that has kind of
inhibited us a little bit, and asecond one would be in a
(09:50):
similar vein.
Like all the bookkeeping andaccounting and everything for
non-profits is much more complexand because it's less common,
it's much harder to findaccountants who understand what
it is they're supposed to do.
You know, if you show up in theUS with a normal C Corp, in the
UK with a normal limitedcompany, any accountant can help
(10:10):
you out and do the normal thingthat they do for all companies.
But when you're a non-profitit's a lot harder to find anyone
who understands what it is thatthey're supposed to do and
you're supposed to do and how itshould all work.
So finding partners andsuppliers on the kind of legal
financial side in general is alot more challenging.
That's probably the biggesthurdle.
Àlex Rodríguez Bacardit (10:30):
Well, one of the good things about
running a non-profit and Icreated one in the past is that
people will help you for freebecause it's a deep belief in
the vision, right?
So that's something that maybenot at the level that I will do
this multi-million financialbookkeeping for you, but maybe
somebody will introduce you tothis other person that will do
this multi-million financialbookkeeping for you, but maybe,
you know, somebody willintroduce you to this other
person that will do freeconsultation and this, or lend
you some sort of services forfree because it's a nonprofit
(10:53):
and therefore you will also getsome kind of I don't know if you
need it, but like donations orsponsorships by companies that
you wouldn't get otherwise,because big corporations most of
the times they have thesebudgets allocated for nonprofits
and social responsibility andstuff like that.
I don't know if that's the case,but if you want to run faster
(11:14):
and again, I don't think this isthe path to follow for Ghost,
but you've gone from under amillion in annual revenue around
2019, 2020, if I remembercorrectly to public numbers, you
posted one month ago 7 millionin annual revenue, right?
So if you really believed inthe long-term vision and
(11:34):
adopting larger and broader userbase by adding more money to
the machine, right?
What financial levers could youuse for that, seeing that loans
not available, vc money notavailable, maybe cash in the
bank?
What else is there?
John O'Nolan (11:55):
There's not much.
It sounds like you've eitherhad better luck than I have or
you have better friends than Ido both as possible.
But no, we've never had anydonations from corporations that
have a budget for giving moneyto nonprofits never once.
And we don't.
We get almost nothing for free.
In fact, we have to pay forpretty much everything.
(12:17):
It's very uncommon that we'llget totally free services.
I think the biggest one isprobably our GitHub account,
where we have, you know, aGitHub pro account that they
give for free to open sourcemaintainers.
But we do get generousdiscounts from some suppliers,
but not as many as you wouldthink.
To some extent, there isn'treally much in the way of levers
(12:39):
that you can pull.
Of course, all investorslooking for some sort of return
on investment right, whichusually means they need to own
part of the company, which meansshares.
Nonprofits don't have shares,so investors are not so
interested.
I do get two or three emailsfrom VCs every single week who
have done enough research to seethat Ghost is a fast growing
(13:00):
tech platform and express theirinterest in investing, but not
enough research to understandthat we're a nonprofit and can't
take investment, which isalways quite amusing.
So I often reply and say we'rea nonprofit but donations are
welcome, and to date I've neverhad a reply beyond that point,
weird right, that's where youlose them, right.
Yeah, exactly, really strange,we also get sorry, go ahead.
(13:24):
There are funding sources for,sorry, there are bodies that do
grant funding in various places.
The highest concentration is inthe US, but we're not a US
company.
Other nonprofits and dophilanthropic work, typically
have either regulation driven orinvestor driven or partner
(13:47):
driven limitations on wherethey're allowed to distribute
funds and that's almost alwaysdomestically.
So there's not an opportunityfor us there right now.
So in some ways it's been anadvantage.
It's been a forcing factor forus to build a sustainable
business, build a revenue streamthat makes us 0% dependent on
any source of outside fundingand make sure that we can
(14:09):
sustain our operationsregardless of anyone else's
support, and that's been a goodthing for the long-term health
of the business.
But it's definitely been alimiting factor.
The speed at which we're ableto move right now is limited by
how many developers we canafford to hire, and that's
limited by how fast revenuegrows, and it's been slow but
consistent growth for about 11years and we just grow at the
(14:32):
pace we're able to.
Àlex Rodríguez Bacardit (14:34):
And actually one funny thing is we
also get VCs trying to invest inMarch-based.
We're a service-based company.
It doesn't really make sense,but I guess that when you
automate your email outrates,that's what happens.
Second, the best investors inthe world are well known for
doing nothing with theirinvestments and just sitting it
out, and so it really makessense.
It would be a very, very goodinvestment, a return on
(14:57):
investment, to invest insomething like sustainable
growth and stuff like that.
Oftentimes and you know it alsovery perfectly is lifestyle
businesses like, for instance,mine and we are regarded as
unambitious, right, that'ssomething I get all the time
when I visit Silicon Valley islike, why don't you do something
more ambitious?
(15:17):
Why don't you raise funds?
Or like merge with anothercompany, start like hiring like
crazy and triple your size everyyear and stuff like that.
It's like I really don't wantto do that.
Like, who can really assure methat I will not grow to 200
people in one year and then goback to 100 and in the next year
(15:38):
?
Right, I don't want to be anaccordion company, and we've
seen plenty of those in theVC-backed world and in companies
relying and believing in hypergrowth.
That's something that we're notreally fond of.
I'm much more leaning towardsthe side of lifestyle and
sustainable and slow growthbusiness.
You know, 20% a year, it'sreally good.
(15:58):
20, 25, 30 on a good year, it'sreally good.
For 10 years now I'll take itany day pay dividends and stuff
like that Team is happy, paygood salaries.
But I think that's not themantra around the tech scene,
that perhaps we have lost thenarrative and I'd like to see
the tables turn.
But I don't think this ishappening anytime soon.
(16:18):
Look, by now I think we've lost100% of the developers that
were listening to the podcasttalking about nonprofits and
financial jargon.
So I wanted to have a couple ofmore technical questions before
delving into the vision and theother topics that I wanted to
cover in the podcast.
(16:38):
More like product and strategy,right, but the technical ones
for the developer bait here isyou chose a tech stack based on
Nodejs.
I don't know if you can tell uswhy and whether that was the
original tech stack or has itchanged and evolved over the
years?
John O'Nolan (16:54):
Yeah, no, that was the original tech stack we
picked.
Nodejs when it was version 0.10back in 2013.
So Node was in its mostly hype,not really proven stage, and
we'd started building the frontend of Ghost in JavaScript point
(17:20):
.
We thought, well, what if thewhole thing was javascript?
And just started exploring thatidea?
The obvious benefit would bethat it would be incredibly fast
, and the tangential benefit,which people used to say a lot
more at the time than they donow, was that you could have
your full stack written in asingle language.
You know that seemed that wasan advantage that was pitched of
nodejs at one point, though Ithink at at this point,
front-end and back-end havesufficiently diverged that it's
(17:41):
almost irrelevant that they'reboth JavaScript, so it was a
little bit of a bet.
It was a punt on a newtechnology that seemed very
promising, and it worked out.
Part of the attraction ofGhostFund we launched was that
it was using this new technology.
It was very, very fast, andthat's what attracted people to
it.
(18:01):
The disadvantage of choosingNodejs was that it was so young.
It was a very, very immatureecosystem and, for the most part
, up until relatively recently,people really weren't building
full-stack web applications withNode, so there was quite a lot
in the way of NPM packages formaking APIs and microservices,
but almost nothing when it cameto building a full web app with
(18:26):
Nodejs, because it just wasn'tthe popular use case.
So most of the types of thingsthat we needed to build for
Ghost we couldn't just pull in adependency or a package for to
help.
We had to write it ourselves wehad expressed, basically give
us roots and views.
But other than that, we kind ofhad to invent everything from
scratch, and that's been a realchallenge over the years.
(18:48):
I think our pace of developmenthas been slowed versus if we
had picked something likeLaravel or Rails that are so
well engineered, so welldesigned for doing exactly what
Ghost does, whereas with Node wekind of had to invent a lot of
our patterns, or essentially allof our patterns, from scratch,
and that's just taken a lot moretime.
Àlex Rodríguez Bacardit (19:06):
But maybe because you had to develop
it yourself, you could continuebeing so opinionated as a
company, like now, that we haveto develop this framework for
doing this exact functionality.
We're going to build it the waywe want it to be right.
I don't know if that's the caseor you have any specific
examples about that.
You would think, but no kind ofthe opposite.
John O'Nolan (19:30):
I think you have to say, yes, it's lovely to
imagine that, if you have theultimate freedom to build
whatever you want, that you willmake better decisions and come
up with a better architecturethan something like Rails or
Laravel, but in reality, we werefocused on delivering the
product to users, and whatevercode needed to be written to
deliver the value to ourcustomers.
That we needed is what is thecode we ended up writing, and
some of that was good code andsome of it is what you would now
describe as technical debt.
(19:50):
It was not good code and wegrew Ghost as a product while
Nodejs was growing as a piece oftechnology, and a lot of the
fundamental building blocks thatare now common in Nodejs-based
apps just simply didn't exist in2013.
So there's a lot of cruft,there's a lot of stuff that we
don't love about our code baseand there's a lot of bits of
(20:12):
frameworks that I really don'tthink anyone's excited about
writing.
You know, like an ORM In Nodejs,orm is like one of the great
unsolved problems.
That isn't a good one.
There's a bunch of old ones,there's a couple that are sort
of some sort of enterprisestartup play, but there isn't an
equivalent to ActiveRecord orEloquent that's well-maintained,
(20:33):
that is performant and that'scompatible with the way most
people build web apps.
So yeah, I don't know, it'sbeen a big disadvantage not
having that.
And now we're kind of 10, 11years later in the place of
trying to refactor and extractsome of the fundamental parts of
the app into something thatresembles more of a framework.
(20:54):
But that's a very slow process.
Àlex Rodríguez Bacardit (20:57):
That is a perfect assist and very
timely.
I was going to go intotechnical depth.
So have you had to rewritecertain parts of the product or
do a full rewrite over thecourse of this decade?
John O'Nolan (21:11):
No, I'm a big believer in a full rewrite is
never the answer.
There's a good Joel Spolskypost on that that I think has
aged very well and continues tobe relevant.
We have had to evolve andupdate parts of the architecture
as we've grown, but we alwaystry our best to do that
incrementally.
I think it rarely makes senseto pursue a refactor project
(21:35):
which doesn't have a clearcustomer-based goal that it's
aiding, just a developer-basedgoal of making the code nicer.
So we try and always focusthings around.
Why do we need to refactor?
Is it because we're blocked andadding a new feature?
Is it because performance hasbecome too slow?
Is it because stability ordowntime is being affected by
this piece of code which ispoorly tested and it's difficult
to test?
And as we do that, we kind ofreplace, upgrade, move around
(22:01):
the pieces of the app that feellike they're causing the most
problems.
But it's always aligned with acustomer problem or a user
problem that we're trying tosolve.
Àlex Rodríguez Bacardit (22:08):
And how do you approach technical debt
in the company?
Now that you mention it, do youhave as a KPI internally?
Some companies have got it as abusiness KPI, even I've heard
recently.
How do you approach it?
John O'Nolan (22:20):
yeah, you know, probably we could do better in
how we approach it.
It's fairly abstract.
We kind of discuss problems asthey come up, we notice and try
and record which parts of theapp, which areas of the app, are
the most common sources offriction within our development
team.
That's our kind of primarybarometer for where the
technical debt is and then wejust try and prioritize it in
(22:44):
the context of what fits withour goals for the quarter
outside of technical debt.
So we're always trying toaddress some of it and pay it
back effectively as we go.
But I don't know if we have areally clear, refined system for
measuring it as a KPI or havingspecific goals around technical
debt itself.
Àlex Rodríguez Bacardit (23:03):
Well, also, your team isn't as weak as
companies that probably havehad to adopt a technical debt as
a KPI because they havethousands of engineers and every
day, between idea and goinglive into production, is
probably millions of dollars inrevenue.
Right so, and your team?
Right now, how many people have?
You got 34, I think 34.
(23:25):
That's great.
I mean, I love seeing the growthof these kind of companies
because it gives me hope thatgood software can still be built
.
And actually, you know, goingback to the bootstrap thing, I
think that I shared the vision.
In our last episode we had GregScown, one of the founders of
TaxExpander, and he went reallydeep in explaining their product
(23:48):
strategy and how they reallypolished and ironed out every
single detail of the application.
And I made the remark that Ilove products that don't change
because it means they work right.
So products like Harvest orTextExpander or Magnet they're
products that are, you know,indie developers have made them,
or they're bootstrap companiesand they do change, but they
(24:10):
change in the background so thatusers don't see it right.
Because Greg mentioned that thiskind of software leaves a lot
on reputation because otherwise,if the users get pissed, they
will go to the next to do thebiggest, the bigger commercial
thing right, and so they have tochange a lot of the background
stuff.
You know so back-end algorithmsand calculations, stuff like
(24:30):
that, not reflecting it on thefront end, and by by doing so
they don't break.
These products don't break.
I've been using now goes forfor two months and three months
now I haven't found any singlepoint of breaking, not a single
bug.
That's very remarkable.
The integrations you knowconnect to Stripe boom done.
(24:51):
Connect to Cloudflare boom done.
Everything works seamlessly.
I love this kind of experience.
I wish more companies wouldadopt that no-transcript
(25:25):
repetitional damage to ghostover the years and the years
that perhaps I missed kind of.
John O'Nolan (25:31):
We had a pretty big fuck up that we sort of
suffered from but it wasn'ttotally our fuck up in mid 2020,
so we were all kind of lockeddown.
Everyone's stressed with covidand what's going to happen in
the world.
There was a there waseffectively a zero-day
vulnerability in salt stack,which we were using to
(25:52):
orchestrate servers and ourhosting platform, and salt stack
did a remarkably horrific jobof to close disclosing that
there was a vulnerability of anykind whatsoever.
I think they, if I remembercorrectly, they sent one slack
message to a private communitythe day before the patch was
released and then just didn'ttell anyone else there was any
(26:14):
vulnerability at all.
And so we discovered that thisvulnerability had been exploited
to install crypto miners on ourapp servers.
Not a good discovery to make.
You try and figure out why yourCPU is like past 100% and
what's going on, only todiscover that someone's trying
to get that sweet sweet Monerofrom your very expensive VPS
(26:43):
Monero from your very expensiveVPS.
And so when we discovered this,we reacted very quickly to both
patch the vulnerability, verifywhether or not it had been
exploited in any other way togain access to our network,
reset, basically wiped allcredentials of every single user
of any kind, from all of ourdatabase so that, on the next
login users be forced to do apassword reset, only to realize
that we didn't have thefunctionality for wiping
(27:06):
credentials and force people todo a password reset.
So, building and deploying thatquickly, and then we disclosed
that the exploit had happened.
Now the problem was this allhappened over a weekend and we
move very quickly because we'rea small team and we're able to
get things resolved very quickly.
It was not only us that hadsuffered this exploit, because
(27:26):
salt stack had done such aterrible job of disclosing it.
Essentially, everyone in theworld using salt stack had been
infected with the same cryptominer, which effectively blatted
the internet looking for openendpoints that it could exploit
and installed itself everywhereit could.
But we were the first todisclose, and so what happened
(27:48):
was the news cycle was Ghost hasbeen hacked for most of the
week.
Not, saltstack has avulnerability.
Not, this is affecting lots andlots of tech companies due to
this underlying piece of opensource technology that lots of
companies are using.
It was us in the headlines, ushaving been hacked, so
(28:08):
reputationally, that was notgood.
I don't know if it probably didcost us some revenue, but it's
always difficult to say how muchright.
But the lesson there was do theright thing but do it in the
right timing.
I think if we had just not beenquite so quick to disclose it
within kind of 24 hours of ithappening, but had just got all
(28:29):
our ducks in a row, goteverything buttoned up, made
sure everything was workingagain, then perhaps wouldn't
have been quite so extremelyfocused on us for a solid week
of the press cycle.
Àlex Rodríguez Bacardit (28:42):
One thorny thing about running
companies that are focused onemployee well-being can work
like Valence and totally goingagainst the hustle and the grind
is that our employees are notsupposed to work on the weekend
or long shifts like in othercompanies.
Right, we're not freakingMcKinsey, we're not Deloitte,
(29:03):
and our employees take their, ofcourse, the natural time off of
the weekends.
And then some employees of ours, at 2 pm, they're done because
they start at 5 am in themorning, right.
But when it comes to working ona weekend for an emergency like
that, what kind of policies doyou have in place to compensate
for that?
How do you rally the troops?
John O'Nolan (29:24):
We have an on-call team and incident management
process.
So essentially, if there's anydowntime or critical emergency,
we have a way of notifying teammembers and for our on-call
engineers infrastructure teamthat's a part of their general
job description.
There's a rota that we havesystems for managing it.
And for other team members, wejust ask people to help when we
(29:45):
need it.
And then on the flip side, onthe kind of work-life balance
side, we actually does afour-day work week and we have
unlimited time off policy.
So we kind of ask everyone tohelp out when we need it and we
also try and give people freedomto help out when we need it and
we also try and give people,you know, freedom and time off
when they need it.
Àlex Rodríguez Bacardit (30:03):
That's amazing.
Have you encountered anyproblem with people?
Because we did have unlimitedtime off in the beginning, but
we found out that people wereactually not taking holidays.
They were not taking days offbecause the founder team we were
not doing it.
So we had to start being reallyresilient and vocal on working
only 40 hours per week even thefounders right Because otherwise
(30:25):
people were not taking holidays.
Did you have an issue with that?
Or have you set a minimum ofdays a year that you have to
take off?
John O'Nolan (30:33):
Yeah, we set a minimum.
We found the same thing.
We set a minimum of two weeksper year that have to be taken
off, and if people haven't takenthat off in December, then we
closed the office at the end ofDecember for the Christmas New
Year period and if peoplehaven't taken it off by then, we
kick them out early.
Effectively, I think we've onlyhad to do that maybe once,
(30:56):
because if you say that thenpeople are more likely to then
just go and book it when it'sconvenient for them.
But yeah, I think, I think it'salways a bit of a thing as
unlimited time off is great, butthen if you can take unlimited
time off, then you you alwaysdefer it to later.
So it was part of themotivation for switching to a
four-day week was to recognizethe kind of intensity of work I
(31:20):
think, especially just inquote-unquote this day and age
being remote, crossing lots oftime zones, having customers
everywhere, having incidentsthat are sometimes out of hours
that people have to deal withand trying to combat that with
more of an intentional,conscious break for people every
single week as opposed to onlywhen they happen to have time
(31:43):
off booked.
Àlex Rodríguez Bacardit (31:45):
One.
I've been trying to do thefour-day work week and it ends
up on.
You know you have meetingsscattered all over the week.
I could concentrate in threedays or four only, and then the
fifth one I would dedicate itonly to admin work in my house,
because everything is soaccumulated and then you take
this day off but it ends up.
You're working on somethingelse.
(32:06):
But is there any otherdisadvantage you've found so far
with the four-day work week?
Because this is an experiment Ithink we ought to do eventually
, being a services company, it'smuch more complicated because
the clients pay for full weeksand they will not understand for
that.
But at the same time I knowother agencies are charging for
full days of work consisting ofsix hours only instead of eight.
(32:29):
So at the end of the day itkind of like evens out.
I don't know, but probably youcan.
You split the team in two andsome of them work one half of
the week, the other team worksthe other half of the week, and
stuff like that.
I don't know what disadvantageshave you found with this new
approach?
John O'Nolan (32:48):
Yeah, I think what you mentioned about having
other work and admin to do isone of the reasons why the
four-day week is valuable.
I think everyone has someversion of of that, whether it's
, you know, appointments orshopping or kids events or just
life maintenance needing to go,and I don't know, do a tax
(33:08):
return, talk to an accountant,figure out something with your
bank account, and then a lot ofthose things can't be done on
weekends because the place thatyou need to do the thing is
closed or you just have run outof hours in the day to get to
the, the life admin that isnecessary of all of us and is
difficult to do alongside afairly intense workflow job.
(33:29):
So we use the.
The four-day week, I think,helps give people space to use
that time where they mightotherwise feel like they're
losing the weekend, which whichreally sucks.
I think that's a big advantageof it.
The most significant downsidethat I think we're still trying
to figure out is the context gapbetween Thursday and Monday is
(33:52):
much larger than you wouldexpect compared to the context
gap between Friday and Monday,if that makes sense.
So having the three days offand returning to whatever it was
you were doing last week isslower to get up to speed and
conversations that start on aThursday difficult to remember
what we're talking about againwhen you get to Monday morning.
Or if something on a Thursdayafternoon is not quite done,
(34:15):
it's very easy to kick it toMonday, but then suddenly you're
at quite a big disadvantageafter that big gap of time has
passed.
So I think figuring out how tooperate efficiently in the four
days that are available for workbecomes much more important.
I don't think we're super goodat that yet, or we haven't
figured it out got the playbookdown just yet but it's one we've
(34:36):
noticed and talked about bookdown just yet, but it's one
we've noticed and talked about.
Àlex Rodríguez Bacardit (34:45):
Also, there is not enough data out
there explaining correlationbetween four-day work weeks and
taking less sick days off orneeding less vacations, and
maybe people who used to takethree weeks off in the summer
break now they take only two.
There's not enough volume ofdata or not enough companies
that have implemented it foryears, so the data set is not
yet relevant, but I assume therewill be something there.
I can think that if I amworking on a four-day work week
(35:09):
for a year, I will not beburning out so easily and maybe
I don't need to take those longbreaks over the summer and stuff
like that.
But I don't know.
That's my observation.
I wanted to circle back to thetransparency topic.
You've done a remarkable job atthe sharing.
(35:30):
The journey goes, the ups anddowns and all the hard things
about running this kind ofbusiness all the hard things
about running this kind ofbusiness.
In the same vein, joel fromBuffer made something really
extraordinary for me a fewmonths back when he shared data
openly about the plateau theyhit and the loss of revenue they
(35:52):
hit one year ago.
Something like that, becauseusually building in public is
great, but if you notice mostpeople building in public, they
only build in public on the wayup.
When they go down, they stopbuilding in public, they stop
sharing metrics because thatprobably causes an accelerated
degrowth.
Perhaps If people see that thebusiness is going down, oh, I
(36:13):
might as well change to anotherplatform.
Right, and Joel shared that, andthat was remarkable because not
a lot of people will admithitting a plateau or their
numbers going down in public,and I think that's an exercise
of responsibility that everybodyshould do.
I understand that publiclytraded companies cannot do that
(36:34):
I mean they have to do that butthey cannot communicate certain
things too early in order not toalter the value of the stock
and alarm the investors andstuff like that.
But more companies that arefinancially independent should
be more open about this kind ofdata.
I found myself that when Ishared some stuff not working or
(36:57):
some problems, we hadinternally some fuck-ups.
I'm very public about ourfuck-ups as a company and not
only we get a lot of engagementin social media.
It's not a metric I really carefor, but we get recognition and
perhaps better contacts.
We're contacted by people like,oh look, you shared that, that
(37:17):
was a fuck up.
It was great.
I learned from that.
I want to work with you guys.
I don't know if you notice likea spike in talent or kind of
like signups every time youshare your data.
John O'Nolan (37:31):
Yeah, particularly the.
You know it's the tried andtrue indie hacker thing.
Quite easy to move the needleon the startup mri chart in the
early days.
You know you, you do one bignew thing and mri doubles or
(38:02):
there's like a 50 change.
I think a lot of people stopsharing revenue, not necessarily
when they plateau, but whenthey just get above a certain
size.
When you pass around a milliondollars a year in revenue,
sharing revenue metrics I thinkstarts to be more of a
disadvantage than an advantage.
At that point.
(38:22):
You've been around forpresumably long enough that
people kind of already know whoyou are within the startup space
and you're not small enoughanymore to be considered an
underdog or like someone to berooting for because they're just
getting going.
But you are starting to becomea target for competitors and
(38:42):
competitors.
Having insight into your dataand your growth rate and what's
working, what's not working asintelligence for what they
should do and how they cancompete with you becomes an
increasing risk.
So I can understand why a lotof startups stop sharing revenue
as they get to a certain size.
At the same time, it's verydifficult to influence those
metrics in a meaningful way.
That tells a good story as youget bigger.
(39:04):
Any changes I make for the nextsix months, I'll be lucky if it
moves the needle by 10%, whichis just not nearly as much of an
interesting story.
But in the context of Ghost, weare very committed to the
transparency piece, in partbecause it aligns so much with
the ethos of the open sourceprojects as well as the
nonprofit foundation.
(39:24):
I think transparency goes handin hand with both of those.
But I do think you have to bemindful of the way in which you
do it and the reasons for whyyou're doing it.
For us, the reasons are thatit's aligned with the project
and the structure.
For other people, the reasonsthey do it are for marketing and
because they want to getattention.
But I don't necessarily thinkeveryone should share metrics
(39:48):
always and it's always good.
I think you should have a goalin mind for why you're doing it
and then measure against it.
But to your point, undoubtedlyit's something that we hear is a
strong signal when it comes tohiring team members, something
that they like, something thatthey appreciate, like the
transparency and the rest of thestructure of Ghost is a huge
pull for us being able toattract talent to work here and
(40:12):
it's a big pull for customerswho have, at various points, had
the rug pulled out from underthem from other tech companies
that start up, shut down.
They have no idea if they'redoing well or poorly, or if
they're running out of money orneed to raise more.
In them being able to trustthat we're going to stick around
, we're doing okay and they caninvest their time and energy
into building on top of Ghostwithout the fear that it's going
(40:35):
to sell to Yahoo next year.
Àlex Rodríguez Bacardit (40:37):
Yeah, exactly, I mean on the building
in public for indie hackers.
There's a lot of upside,there's literally no downside,
to saying, oh, we're doing 2000euros in MRR in our first month
or year, something like that.
There's no downside to that.
I have never set the limit fromwhich it is dangerous or
(40:59):
calculated risk.
One million in AR makes sense.
For another reason that doesn'tapply to you guys, which is you
will never sell your company.
But when you're sharing yourdata publicly, you're also
sharing your valuation, yourpotential sale price, right.
So that's some kind of numbersthey can use to negotiate
(41:19):
against your asking price, andso it doesn't apply to you.
But you have to take it intoconsideration if you have
another kind of companystructure.
The other question I have foryou with regards to
communication and style becausethis is one of my personal
struggles is by having thistransparency core value in the
(41:40):
company.
Where do you set the line intransparency right?
Over-communication versusunder-communication?
What's the right timing tocommunicate?
Oh, look, guys, in three monthswe'll have no projects for
everybody.
You better start looking foranother job or you know you're
going to overcome that situation.
So you just don't sharebeforehand, so that you don't
create an internal sense ofalarm and an emergency in the
(42:04):
company, right how to set aprotocol for that and a standard
in communication, especiallynot of bad things, but of
preparing people for potentialbad things that might happen.
You know, when there was COVIDhit, potentially we could lose
(42:24):
all our clients Didn't happen,but it could have happened.
Other companies shut down.
In the last 24 months, mostboutique agencies like ours have
either shut down or sold toother companies at a discount
because they were doing verypoorly, and so our employees
shared the concern like, hey,everybody is doing poorly, how
are we doing?
We're doing great.
(42:44):
But I have to let you know thatthis is happening out there.
It's very cold out there,salaries are going down, there's
a lot of layoff season andstuff like that.
But I don't know Like I don'tknow if you've got the recipe
for that or, because you lookvery calm and very you know very
thoughtful about everything yousay, it looks like it's pretty
(43:05):
calculated every time you shareit, but I don't know if you also
internally struggle with this.
John O'Nolan (43:12):
Yeah, I guess I kind of think about this in the
context of we try and share asmuch data with the team as
possible, particularly data thatwill help them be able to do
their jobs more easily.
So on day one when someonejoins ghost, they get full,
unlimited access to revenueanalytics that they can dig
(43:33):
through the entire thing, everysingle metric that we have
performance over the last 10years, and see how the company
has performed.
And it serves two functions.
One, it kind of sets everyone'smind at ease about how we're
doing, like, how are thingsgoing?
Are we flatlining, are we goingdown, are we going up?
You know, you just take awaythat whole category of questions
Because in reality, if it'stough, if your team are worried
(43:55):
about that, if they're thinkingabout that, then it's going to
be difficult for them to beproductive on what you hope it
is that they're working on.
So you potentially could bewasting energy in that area when
you don't need to.
Another thing is they can seethe effects of their work.
You know, if they launchsomething or work on something,
it can see like did the numbersgo up, did they stay the same?
(44:16):
Or if they have a theory abouta segment of customers like we
think business customers, whoare all asking for this new
feature, are growing faster thanthe other customers and we need
to devote more resources tothem.
If they have access to the datato go and validate that and
look at the segment of thosecustomers and how they're
performing and what they'redoing, then they can make better
decisions in productdevelopment.
So we try and open up as manythings as we can unless there's
(44:40):
a really clear reason not to.
We essentially default to.
Everyone can see everything,unless there's a particular
reason of some kind legalcompliance or otherwise that
they shouldn't be able to.
And then when it comes tosignaling you know, signaling
bad news or talking aboutpotential disadvantages that are
coming up we do our best as the.
(45:01):
When the team was really small,we did, you know, a weekly all
hands and we just talked openlyabout everything with everyone.
Now we have, you know, amultitude of teams working on
different things, differentpriorities.
You a weekly cadence wouldprobably be more distracting
than helpful.
So we do a big all-hands everyquarter now, sorry, a big
(45:22):
all-hands every month.
And then that includes astrategic kind of planning
session and bigger picture viewevery quarter and that's the
place where we bring up anythings that, as a company, we
feel like we're strugglingstruggling with risks that are
potentially on the horizon andthen open it up for anyone who
has feedback or questions totalk about it.
And and then we do the samething in person once a year on
(45:43):
our team meetups, which isalways a good venue for bringing
if there are difficult topicsthat need to come up around
company performance or like howwe're doing or whether we're
working well together or there'sroom for improvement or people
are frustrated, perhaps withprocesses or systems that aren't
good.
You know, the one of thebiggest downsides of remote is
(46:03):
empathy doesn't translateparticularly well through 2d
zoom calls.
Being in person makes it a loteasier to have those
conversations.
There's just much more of ahuman element to it.
So we try and always createspace for those things to come
up when we're sitting together,hanging out in person, and that
always is a good reset or a goodplace to do those types of
(46:27):
things.
Àlex Rodríguez Bacardit (46:28):
Good.
And to wrap up the topic oftransparency, I wanted to ask
because, before you share a techfuck-up that was probably
somebody else's fault, do youhave any personal fuck-up that
you've done in the company thathas resulted in some kind of
like financial?
In other words, what's beenyour most expensive fuck-up at
Ghost?
(46:48):
It has to be yours.
John O'Nolan (46:51):
Yeah, I've had plenty.
Àlex Rodríguez Bacardit (46:53):
It's a signature question of the
podcast, so yeah, I've had tons.
John O'Nolan (46:58):
It's just difficult to pick one.
I don't know like, yeah, youcould probably tie this back to
revenue somehow, but it's it'sdifficult one.
I once managed to get ourentire website banned from
google, for it might have beentwo months, oh wow, which is
pretty, pretty fucking bad.
So the what happened was wewere getting inundated with spam
(47:19):
.
So these and this enterprisingseo agency in malaysia had
decided that they were going toexploit ghost's free trial
system so you could sign up fora free trial which would give
you a website, and then theywould put loads of porn the
website and try and use that forSEO.
(47:39):
I'm not entirely sure what themotivation was, but it was.
Maybe it was an SEO agencyworking for a porn company.
It's not entirely clear.
In any event, we were like afive person company.
I think this was 2014 or 15.
And the resources that they wereusing up with this spam, this
automated spam, was intense,like it was costing money.
It was making servers go downbecause of the level to which
(48:02):
they were doing it, and so,eventually, how did I do it?
I found through some sourcecode.
I can't remember.
I found the agency in malaysia.
I called them up 2 am in themorning.
That time I clearly worked somedude up, presumably the owner
of the agency, and I said please, can you stop spamming us?
We, we need this to end.
(48:24):
And I'm trying to remember thesequence of events here.
I think he sort of said hewould and then he didn't.
In any event, we had a biginflux of requests, spam
requests, and again one day, andwe were going through the
server logs, I was lookingthrough Cloudflare and there was
a big influx from this IP.
That was Google, you know, likeGoogle Cloud Platform, so it
(48:44):
blocks that whole IP range likeno more VPS spamming our servers
for you.
Anyways, long story short, thatwasn't Google Cloud Platform,
it was Google Web Crawler tryingto index our site who had just
been banned from absolutelyeverything.
It wasn't until, I don't know,a month later or something, when
I Googled something in our docs, that nothing came up.
(49:04):
That's strange, it always comesup.
And then I realized that noneof ghostorg was coming up, and
then I realized that we were nolonger ranking for a single
thing whatsoever.
And that was entirely on me anda hasty click in a Cloudflare
dashboard.
Fortunately it recovered justfine, great.
Àlex Rodríguez Bacardit (49:23):
Well, thank you for doing this,
because actually we need moreexamples of that.
People think that othercompanies work perfectly and
seamlessly, while our personalprojects are always like shit,
but I think we make a prettygood job at democratizing
failure in this sense.
Look, I was worried because Ihad another block of topics that
was how to more like productstrategy and stuff.
(49:44):
We will not be covering thisbecause we're running out of
time, and one of the keyquestions there was, like I
haven't found the formula tosort of resell Ghost as an
agency, because for me it lookslike a very good platform to
build projects on.
But there are certain thingsthat you guys are so opinionated
on like the multi-languagesupport, natively or not adding
(50:04):
content blocks and stuff likethat that other platforms use it
, and you guys have got verylegit reasons not to give them
out of the box, and so I wasgoing to go into that.
But I think from this answeryou've just given me, I've
inferred how not to use Ghost,which I think it's as valuable,
if not more so.
Well, let's just wrap it here,because you've been extremely
(50:26):
generous with your time.
I'm rolling out the carpet foryou.
You got one minute left to sayhow we can help you, how we can
help ghost and any announcementsor anything you want to share
with our community love the hotones reference yeah, if you want
to check out to get it?
yes, I would.
John O'Nolan (50:47):
if you publish online and you have an email
list or you have thoughts toshare with the world, check out
ghostorg.
We're doing a lot ofinteresting things.
The most interesting of whichthat is coming up, that is kind
of on the horizon, is activitypub-based federation.
So soon all ghost sites in theworld will be networked with one
another through decentralizedprotocols that will give you a
(51:09):
social graph inside anindependent website, so you'll
be able to like reply andinteract with other publishers
and audience out there in thebig wide social web, which is
very exciting.
We've been working on this inbeta for a few months and it
completely changes the shape ofthe product in a way that I
think in the next year or two isgoing to be pretty
(51:32):
transformational.
So that's one of the fun thingswe're working on and, yeah,
love it.
If anyone wants to check outthe platform, shoot me a message
on any of the social platformsat Jono Nolan If you have
questions or want to chat.
Àlex Rodríguez Bacardit (51:44):
I totally endorse that as a
personal user and before you go,you got it coming.
Favorite Hot Ones episode sofar.
John O'Nolan (51:51):
I mean, the DJ Khaled's always a classic right
so far.
I mean the dj khaled's always aclassic right, just for for its
incredible failure.
What was the one I watched justlast week?
Oh, ryan reynolds, that one wasreally good yeah, that's,
that's very good there's a few.
I'm struggling to think of themall on the spot.
Jimmy fallon is the one I justwatched.
He was very funny.
(52:12):
What's yours?
Àlex Rodríguez Bacard (52:14):
Shaquille O'Neal, probably Like I don't
understand most of the timewe're speaking, but the faces
and the entire performance issuperb.
That for me is like an all-timegreat.
John O'Nolan (52:25):
Ricky Gervais.
The Gordon Ramsay one was goodas well, just because of.
Gordon Ramsay because of howangry he got throughout the
whole thing.
Àlex Rodríguez Bacardit (52:35):
Like why angry you got throughout the
whole thing, like why do you dothis exactly?
Exactly they're.
They're really good ones.
And then I'm really surprisedby people who you wouldn't
expect it but they toleratespiciness so well, like charlize
their own, for instance and I'msurprised gary venetro.
Yeah yeah, exactly, some peopleare like yeah, yeah, spicy, but
you know, I'm like yeah, holyfuck like I.
I have the bomb, the bomb, Ihave it in my fridge.
(52:56):
It's like literally unbearablestuff like that I love spicy,
and that's why I buy all ofthese terrible sauces.
And when I see some of theguests that they don't break a
sweat, I'm like, no, I don'tknow how they can do it.
I don't know how they can do it, but anyways, I don't know how
they can do it.
I don't know how they can do it, but anyways.
John O'Nolan (53:16):
I'm a big Hot.
Àlex Rodríguez Bacardit (53:16):
Sauce fan.
Yeah, love having you on andcongratulations on what you've
built.
Much success for the future.
And also thanks for getting theHot Wins reference first First
person in 150 episodes.
So far that gets it.
John O'Nolan (53:33):
Amazing, I'll take that, thank you.
Àlex Rodríguez Bacardit (53:36):
Thank you very much.
Hello everybody.
I'm Alex, founder of MarsBased,and in this episode we bring
you John O'Nolan from Ghost, acompany behind the popular open
source newsletter and bloggingplatform that we have used here
and there at Marsbase.
I have used extensively as mypersonal website over the years.
My founder to foundernewsletter nowadays is built on
Ghost.
Ghost and with John wediscussed many things that we
(00:23):
have got in common as a companythe transparency, the core
values, remote work,bootstrapping, but also some
particularities they've got.
They are a foundation.
It's a kind of company thatcannot be sold.
They will not be able to takethe usual investment that other
companies might be entitled to.
We speak about the future.
(00:44):
We also talk about theirtechnological stack and we dive
a little bit deep into thetechnical intricacies of their
product, particularly regardingrewrites, technical depth and so
on.
And the question of the day forthis episode will be what
newsletter platform do you usefor your company?
John O'Nolan (01:06):
At.
Àlex Rodríguez Bacard (01:07):
MarsBased .
We're considering switchingover from MailChimp to something
else, and your answers will beparticularly helpful, and so,
without further ado, let's jumpright into this episode.
Good, I think this is recording.
You don't look very nervous.
How are you doing?
John O'Nolan (01:23):
John, welcome to the show.
Thank you for having me seemslike you're used to this.
Àlex Rodríguez Bacardit (01:28):
Not my first podcast, but I'm sure it
will be the best you look moreof a professional podcast
recorder than I do, so it seemslike you're very used to this.
I think it's part of the remotesetup, right?
John O'Nolan (01:41):
yeah, that's right .
After a while you get used tosetup right.
Yeah, that's right.
After a while you get used totrying to have a nice setup and
making video calls feel good.
Àlex Rodríguez Bacardit (01:50):
I wish my team could hear this so that
they also bought the nicecameras and mics and they don't
have to be using these setupsthat introduce a lot of echo
into the calls and stuff likethat Good.
I have so many questions.
I've been using Ghost for 10years, on and off.
I have to say that my journeyhas been interesting with the
(02:10):
platform.
I remember trying it too early,but more recently I have
rediscovered it as a personaluser.
I recreated my website Everytwo years.
I've got this thing where Irebuilt my personal website with
a different framework and I waseager to do it with ghost this
time around.
I have to say there's somethingin it that gave me the wow
(02:33):
impression every step of the way.
I think you guys nail it, andso here goes my gratitude and my
you know I'm super impressed byit.
How much of this is affected bythe fact that you're a bootstrap
company, independent company,and that you can take however
long it takes to polish all thedetails.
John O'Nolan (02:54):
Well, I think it is affected by us being a
bootstraps company, but notnecessarily because we can take
all the time to polish thedetails.
You know we have a lot ofcompetitors, pretty big ones I
think the smallest has $50million of funding and the
largest has $1.1 billion offunding and people, as in end
(03:14):
users, don't particularly careif we're bootstrapped and
independent.
They just want to use the bestproduct right.
So if we took all the time inthe world to just polish
everything and make it perfect,we simply wouldn't be able to
keep up with the expectationsthat people have from products
like Ghost.
But that said, I think beingindependent and bootstrapped
allows us to have more focus,allows us to optimize for
(03:36):
different things, allows us totake different strategic moves
than we might otherwise if wehad investors who were pushing
for growth targets Exactly.
Àlex Rodríguez Bacardit (03:47):
I take it that companies like Basecamp,
ghost, buffer you know thatyou're more well known for being
highly opinionated,bootstrapped or
semi-bootstrapped.
You can take more time topolicies, and I know you're
right.
I grant you that you don't haveall the time in the world.
However, you are aware of asubset of users that they will
(04:09):
buy your product, whatever ittakes they don't really care
about.
Maybe there's a bettercompetitor, maybe there's a
cheaper alternative, but I willbuy your product because of your
values, your vision, and I'mtotally aligned with your
product.
For instance, in our companyvalues, your vision and I'm
totally aligned with yourproduct.
Like, for instance, in ourcompany MarsBased.
We created the company becausewe read the books of Basecamp
(04:29):
and so for 10 years we werepaying for that product and we
knew there was better softwareout there.
But we said, like, thanks toRails, thanks to Basecamp, we're
here and now right.
So it took really long for usto kind of like, get rid of
Basecamp when it was, you know,unavoidable that we had to use
other products.
But I find myself usingproducts because I really
(04:50):
believe in the vision, right?
So have you calculated how muchof a percentage of your user
base is this fanatical user basethat will pay for your product.
John O'Nolan (05:02):
you know, whatever you do, or not really Haven't
calculated the percentage.
I think that might be quitedifficult to do.
There's definitely some peoplewho are very passionate about
Ghost based on structure andvalues, and that's a really
important part of how we dothings.
I think the killer combination,though, is if you can combine
the two.
If you can have great structureand values as well as a great
(05:22):
product, I think then you'rereally onto something right.
I think of products outside orcompanies outside of the tech
space with things like thisright.
So, like Patagonia is theclassic example, people are
aligned with their values.
People buy Patagonia stuffbecause of their values, but if
they made crappy t-shirts thatfell apart after a while, then
(05:43):
it wouldn't be what it is today.
It's the quality and the carethat goes into the product,
combined with the values.
That, I think, is where themagic can happen, if you get it
right, exactly, you mentioned it.
Àlex Rodríguez Bacardit (05:55):
You created a brand, and one of the
things about creating a brand ispeople will follow it more
strongly than anything else.
People will be or a user basewill be less volatile, if you
will, but that doesn't happen onday one.
That doesn't happen on day 10.
Are you aware of the day wherethat happened, when you made the
click and you started gettingit right, everything you made,
(06:16):
and you started attracting thissubset of users, the cult-like
users that we were talking aboutbefore?
Are you aware of the precisemoment or movement you did in
the company that kind of, likeyou know, turned the direction
of the company into the rightway?
John O'Nolan (06:32):
Yeah, I think Ghosts is an unusual case where
it did happen on day one, andthat's actually to do with the
origin story of the company,which kind of started as just
this blog post pitch of what Iwanted to exist in the world and
the values that I thought itshould have, and so the values
for the company existed beforethe product, before the company,
(06:54):
they existed as an idea and aconcept that I put out into the
world, and so it was actuallythat that captured people's
attention as much as the ideafor the product itself.
So, as we went from blog poststo Kickstarter to the first
launch, the consistent storythat came along with Ghost was
that we were doing this for love, not for profit, doing this for
(07:15):
values, not for growth, andthat was a huge part in the
launch and then also a huge partin the continued growth that
came afterwards.
Àlex Rodríguez Bacardit (07:24):
Exactly .
One of the things that reallyresonates with me is it's very
similar to Basecams until theend of the internet, right, the
fact that you are a foundation,I think, is something that
stands out.
You are a different kind ofcompany, not only legally, but
also existentially.
If you create a company thatyou cannot sell, you will not do
things that will drive youtowards selling the company,
(07:47):
right, and so you operatedifferently.
Beyond the obvious consequencesof running a foundation, that
might not be obvious for 99% ofthe users of the internet, but
you know, in terms of governance, in terms of leadership.
In terms of leadership, haveyou found, for instance, any
restrictions of things youwanted to do, something that has
(08:08):
held you back because you werea foundation, perhaps, like you
wanted to run faster, which is Idon't believe it's in your
ethos as a company, but whatthings did you find out along
the way that were ah, this isholding me back.
John O'Nolan (08:22):
Yeah, there's a bunch of weird non-profit
regulations and things that youget surprised by, and every
country kind of has their ownset of them, so they'll be
slightly different depending onwhere you are in the world.
The one that comes to mindalways immediately is if you're
incorporated as a non-profitfoundation, you are
(08:42):
automatically classified bybanks and financial institutions
as high risk which is fairlylogical, you know, like most
non-profits are dependent onoutside funding, typically don't
sustain their operationswithout some kind of outside
help.
So do you really want to lendmoney to an organization that
may be not able to pay it back?
As a bank, probably not, butit's a blanket rule.
(09:06):
So, for example, a common thingthat software startups do is
they put their hosting billswhich are usually quite
extensive on American Expresscards, and the benefit of that
is you get a delayed timeframefor paying your hosting bills.
But you also rack up a hugenumber of air miles and points
and things that you can useelsewhere.
Amex won't talk to us.
They say oh, you're anon-profit, no, you're
(09:29):
disqualified.
And same with banks.
We haven't ever needed lendingfacilities, but if we did,
they're not available.
Banks won't give you any linesof credit as a non-profit
organization.
So the flexibility there issomething that has kind of
inhibited us a little bit, and asecond one would be in a
(09:50):
similar vein.
Like all the bookkeeping andaccounting and everything for
non-profits is much more complexand because it's less common,
it's much harder to findaccountants who understand what
it is they're supposed to do.
You know, if you show up in theUS with a normal C Corp, in the
UK with a normal limitedcompany, any accountant can help
(10:10):
you out and do the normal thingthat they do for all companies.
But when you're a non-profitit's a lot harder to find anyone
who understands what it is thatthey're supposed to do and
you're supposed to do and how itshould all work.
So finding partners andsuppliers on the kind of legal
financial side in general is alot more challenging.
That's probably the biggesthurdle.
Àlex Rodríguez Bacardit (10:30):
Well, one of the good things about
running a non-profit and Icreated one in the past is that
people will help you for freebecause it's a deep belief in
the vision, right?
So that's something that maybenot at the level that I will do
this multi-million financialbookkeeping for you, but maybe
somebody will introduce you tothis other person that will do
this multi-million financialbookkeeping for you, but maybe,
you know, somebody willintroduce you to this other
person that will do freeconsultation and this, or lend
you some sort of services forfree because it's a nonprofit
(10:53):
and therefore you will also getsome kind of I don't know if you
need it, but like donations orsponsorships by companies that
you wouldn't get otherwise,because big corporations most of
the times they have thesebudgets allocated for nonprofits
and social responsibility andstuff like that.
I don't know if that's the case,but if you want to run faster
(11:14):
and again, I don't think this isthe path to follow for Ghost,
but you've gone from under amillion in annual revenue around
2019, 2020, if I remembercorrectly to public numbers, you
posted one month ago 7 millionin annual revenue, right?
So if you really believed inthe long-term vision and
(11:34):
adopting larger and broader userbase by adding more money to
the machine, right?
What financial levers could youuse for that, seeing that loans
not available, vc money notavailable, maybe cash in the
bank?
What else is there?
John O'Nolan (11:55):
There's not much.
It sounds like you've eitherhad better luck than I have or
you have better friends than Ido both as possible.
But no, we've never had anydonations from corporations that
have a budget for giving moneyto nonprofits never once.
And we don't.
We get almost nothing for free.
In fact, we have to pay forpretty much everything.
(12:17):
It's very uncommon that we'llget totally free services.
I think the biggest one isprobably our GitHub account,
where we have, you know, aGitHub pro account that they
give for free to open sourcemaintainers.
But we do get generousdiscounts from some suppliers,
but not as many as you wouldthink.
To some extent, there isn'treally much in the way of levers
(12:39):
that you can pull.
Of course, all investorslooking for some sort of return
on investment right, whichusually means they need to own
part of the company, which meansshares.
Nonprofits don't have shares,so investors are not so
interested.
I do get two or three emailsfrom VCs every single week who
have done enough research to seethat Ghost is a fast growing
(13:00):
tech platform and express theirinterest in investing, but not
enough research to understandthat we're a nonprofit and can't
take investment, which isalways quite amusing.
So I often reply and say we'rea nonprofit but donations are
welcome, and to date I've neverhad a reply beyond that point,
weird right, that's where youlose them, right.
Yeah, exactly, really strange,we also get sorry, go ahead.
(13:24):
There are funding sources for,sorry, there are bodies that do
grant funding in various places.
The highest concentration is inthe US, but we're not a US
company.
Other nonprofits and dophilanthropic work, typically
have either regulation driven orinvestor driven or partner
(13:47):
driven limitations on wherethey're allowed to distribute
funds and that's almost alwaysdomestically.
So there's not an opportunityfor us there right now.
So in some ways it's been anadvantage.
It's been a forcing factor forus to build a sustainable
business, build a revenue streamthat makes us 0% dependent on
any source of outside fundingand make sure that we can
(14:09):
sustain our operationsregardless of anyone else's
support, and that's been a goodthing for the long-term health
of the business.
But it's definitely been alimiting factor.
The speed at which we're ableto move right now is limited by
how many developers we canafford to hire, and that's
limited by how fast revenuegrows, and it's been slow but
consistent growth for about 11years and we just grow at the
(14:32):
pace we're able to.
Àlex Rodríguez Bacardit (14:34):
And actually one funny thing is we
also get VCs trying to invest inMarch-based.
We're a service-based company.
It doesn't really make sense,but I guess that when you
automate your email outrates,that's what happens.
Second, the best investors inthe world are well known for
doing nothing with theirinvestments and just sitting it
out, and so it really makessense.
It would be a very, very goodinvestment, a return on
(14:57):
investment, to invest insomething like sustainable
growth and stuff like that.
Oftentimes and you know it alsovery perfectly is lifestyle
businesses like, for instance,mine and we are regarded as
unambitious, right, that'ssomething I get all the time
when I visit Silicon Valley islike, why don't you do something
more ambitious?
(15:17):
Why don't you raise funds?
Or like merge with anothercompany, start like hiring like
crazy and triple your size everyyear and stuff like that.
It's like I really don't wantto do that.
Like, who can really assure methat I will not grow to 200
people in one year and then goback to 100 and in the next year
(15:38):
?
Right, I don't want to be anaccordion company, and we've
seen plenty of those in theVC-backed world and in companies
relying and believing in hypergrowth.
That's something that we're notreally fond of.
I'm much more leaning towardsthe side of lifestyle and
sustainable and slow growthbusiness.
You know, 20% a year, it'sreally good.
(15:58):
20, 25, 30 on a good year, it'sreally good.
For 10 years now I'll take itany day pay dividends and stuff
like that Team is happy, paygood salaries.
But I think that's not themantra around the tech scene,
that perhaps we have lost thenarrative and I'd like to see
the tables turn.
But I don't think this ishappening anytime soon.
(16:18):
Look, by now I think we've lost100% of the developers that
were listening to the podcasttalking about nonprofits and
financial jargon.
So I wanted to have a couple ofmore technical questions before
delving into the vision and theother topics that I wanted to
cover in the podcast.
(16:38):
More like product and strategy,right, but the technical ones
for the developer bait here isyou chose a tech stack based on
Nodejs.
I don't know if you can tell uswhy and whether that was the
original tech stack or has itchanged and evolved over the
years?
John O'Nolan (16:54):
Yeah, no, that was the original tech stack we
picked.
Nodejs when it was version 0.10back in 2013.
So Node was in its mostly hype,not really proven stage, and
we'd started building the frontend of Ghost in JavaScript point
(17:20):
.
We thought, well, what if thewhole thing was javascript?
And just started exploring thatidea?
The obvious benefit would bethat it would be incredibly fast
, and the tangential benefit,which people used to say a lot
more at the time than they donow, was that you could have
your full stack written in asingle language.
You know that seemed that wasan advantage that was pitched of
nodejs at one point, though Ithink at at this point,
front-end and back-end havesufficiently diverged that it's
(17:41):
almost irrelevant that they'reboth JavaScript, so it was a
little bit of a bet.
It was a punt on a newtechnology that seemed very
promising, and it worked out.
Part of the attraction ofGhostFund we launched was that
it was using this new technology.
It was very, very fast, andthat's what attracted people to
it.
(18:01):
The disadvantage of choosingNodejs was that it was so young.
It was a very, very immatureecosystem and, for the most part
, up until relatively recently,people really weren't building
full-stack web applications withNode, so there was quite a lot
in the way of NPM packages formaking APIs and microservices,
but almost nothing when it cameto building a full web app with
(18:26):
Nodejs, because it just wasn'tthe popular use case.
So most of the types of thingsthat we needed to build for
Ghost we couldn't just pull in adependency or a package for to
help.
We had to write it ourselves wehad expressed, basically give
us roots and views.
But other than that, we kind ofhad to invent everything from
scratch, and that's been a realchallenge over the years.
(18:48):
I think our pace of developmenthas been slowed versus if we
had picked something likeLaravel or Rails that are so
well engineered, so welldesigned for doing exactly what
Ghost does, whereas with Node wekind of had to invent a lot of
our patterns, or essentially allof our patterns, from scratch,
and that's just taken a lot moretime.
Àlex Rodríguez Bacardit (19:06):
But maybe because you had to develop
it yourself, you could continuebeing so opinionated as a
company, like now, that we haveto develop this framework for
doing this exact functionality.
We're going to build it the waywe want it to be right.
I don't know if that's the caseor you have any specific
examples about that.
You would think, but no kind ofthe opposite.
John O'Nolan (19:30):
I think you have to say, yes, it's lovely to
imagine that, if you have theultimate freedom to build
whatever you want, that you willmake better decisions and come
up with a better architecturethan something like Rails or
Laravel, but in reality, we werefocused on delivering the
product to users, and whatevercode needed to be written to
deliver the value to ourcustomers.
That we needed is what is thecode we ended up writing, and
some of that was good code andsome of it is what you would now
describe as technical debt.
(19:50):
It was not good code and wegrew Ghost as a product while
Nodejs was growing as a piece oftechnology, and a lot of the
fundamental building blocks thatare now common in Nodejs-based
apps just simply didn't exist in2013.
So there's a lot of cruft,there's a lot of stuff that we
don't love about our code baseand there's a lot of bits of
(20:12):
frameworks that I really don'tthink anyone's excited about
writing.
You know, like an ORM In Nodejs,orm is like one of the great
unsolved problems.
That isn't a good one.
There's a bunch of old ones,there's a couple that are sort
of some sort of enterprisestartup play, but there isn't an
equivalent to ActiveRecord orEloquent that's well-maintained,
(20:33):
that is performant and that'scompatible with the way most
people build web apps.
So yeah, I don't know, it'sbeen a big disadvantage not
having that.
And now we're kind of 10, 11years later in the place of
trying to refactor and extractsome of the fundamental parts of
the app into something thatresembles more of a framework.
(20:54):
But that's a very slow process.
Àlex Rodríguez Bacardit (20:57):
That is a perfect assist and very
timely.
I was going to go intotechnical depth.
So have you had to rewritecertain parts of the product or
do a full rewrite over thecourse of this decade?
John O'Nolan (21:11):
No, I'm a big believer in a full rewrite is
never the answer.
There's a good Joel Spolskypost on that that I think has
aged very well and continues tobe relevant.
We have had to evolve andupdate parts of the architecture
as we've grown, but we alwaystry our best to do that
incrementally.
I think it rarely makes senseto pursue a refactor project
(21:35):
which doesn't have a clearcustomer-based goal that it's
aiding, just a developer-basedgoal of making the code nicer.
So we try and always focusthings around.
Why do we need to refactor?
Is it because we're blocked andadding a new feature?
Is it because performance hasbecome too slow?
Is it because stability ordowntime is being affected by
this piece of code which ispoorly tested and it's difficult
to test?
And as we do that, we kind ofreplace, upgrade, move around
(22:01):
the pieces of the app that feellike they're causing the most
problems.
But it's always aligned with acustomer problem or a user
problem that we're trying tosolve.
Àlex Rodríguez Bacardit (22:08):
And how do you approach technical debt
in the company?
Now that you mention it, do youhave as a KPI internally?
Some companies have got it as abusiness KPI, even I've heard
recently.
How do you approach it?
John O'Nolan (22:20):
yeah, you know, probably we could do better in
how we approach it.
It's fairly abstract.
We kind of discuss problems asthey come up, we notice and try
and record which parts of theapp, which areas of the app, are
the most common sources offriction within our development
team.
That's our kind of primarybarometer for where the
technical debt is and then wejust try and prioritize it in
(22:44):
the context of what fits withour goals for the quarter
outside of technical debt.
So we're always trying toaddress some of it and pay it
back effectively as we go.
But I don't know if we have areally clear, refined system for
measuring it as a KPI or havingspecific goals around technical
debt itself.
Àlex Rodríguez Bacardit (23:03):
Well, also, your team isn't as weak as
companies that probably havehad to adopt a technical debt as
a KPI because they havethousands of engineers and every
day, between idea and goinglive into production, is
probably millions of dollars inrevenue.
Right so, and your team?
Right now, how many people have?
You got 34, I think 34.
(23:25):
That's great.
I mean, I love seeing the growthof these kind of companies
because it gives me hope thatgood software can still be built
.
And actually, you know, goingback to the bootstrap thing, I
think that I shared the vision.
In our last episode we had GregScown, one of the founders of
TaxExpander, and he went reallydeep in explaining their product
(23:48):
strategy and how they reallypolished and ironed out every
single detail of the application.
And I made the remark that Ilove products that don't change
because it means they work right.
So products like Harvest orTextExpander or Magnet they're
products that are, you know,indie developers have made them,
or they're bootstrap companiesand they do change, but they
(24:10):
change in the background so thatusers don't see it right.
Because Greg mentioned that thiskind of software leaves a lot
on reputation because otherwise,if the users get pissed, they
will go to the next to do thebiggest, the bigger commercial
thing right, and so they have tochange a lot of the background
stuff.
You know so back-end algorithmsand calculations, stuff like
(24:30):
that, not reflecting it on thefront end, and by by doing so
they don't break.
These products don't break.
I've been using now goes forfor two months and three months
now I haven't found any singlepoint of breaking, not a single
bug.
That's very remarkable.
The integrations you knowconnect to Stripe boom done.
(24:51):
Connect to Cloudflare boom done.
Everything works seamlessly.
I love this kind of experience.
I wish more companies wouldadopt that no-transcript
(25:25):
repetitional damage to ghostover the years and the years
that perhaps I missed kind of.
John O'Nolan (25:31):
We had a pretty big fuck up that we sort of
suffered from but it wasn'ttotally our fuck up in mid 2020,
so we were all kind of lockeddown.
Everyone's stressed with covidand what's going to happen in
the world.
There was a there waseffectively a zero-day
vulnerability in salt stack,which we were using to
(25:52):
orchestrate servers and ourhosting platform, and salt stack
did a remarkably horrific jobof to close disclosing that
there was a vulnerability of anykind whatsoever.
I think they, if I remembercorrectly, they sent one slack
message to a private communitythe day before the patch was
released and then just didn'ttell anyone else there was any
(26:14):
vulnerability at all.
And so we discovered that thisvulnerability had been exploited
to install crypto miners on ourapp servers.
Not a good discovery to make.
You try and figure out why yourCPU is like past 100% and
what's going on, only todiscover that someone's trying
to get that sweet sweet Monerofrom your very expensive VPS
(26:43):
Monero from your very expensiveVPS.
And so when we discovered this,we reacted very quickly to both
patch the vulnerability, verifywhether or not it had been
exploited in any other way togain access to our network,
reset, basically wiped allcredentials of every single user
of any kind, from all of ourdatabase so that, on the next
login users be forced to do apassword reset, only to realize
that we didn't have thefunctionality for wiping
(27:06):
credentials and force people todo a password reset.
So, building and deploying thatquickly, and then we disclosed
that the exploit had happened.
Now the problem was this allhappened over a weekend and we
move very quickly because we'rea small team and we're able to
get things resolved very quickly.
It was not only us that hadsuffered this exploit, because
(27:26):
salt stack had done such aterrible job of disclosing it.
Essentially, everyone in theworld using salt stack had been
infected with the same cryptominer, which effectively blatted
the internet looking for openendpoints that it could exploit
and installed itself everywhereit could.
But we were the first todisclose, and so what happened
(27:48):
was the news cycle was Ghost hasbeen hacked for most of the
week.
Not, saltstack has avulnerability.
Not, this is affecting lots andlots of tech companies due to
this underlying piece of opensource technology that lots of
companies are using.
It was us in the headlines, ushaving been hacked, so
(28:08):
reputationally, that was notgood.
I don't know if it probably didcost us some revenue, but it's
always difficult to say how muchright.
But the lesson there was do theright thing but do it in the
right timing.
I think if we had just not beenquite so quick to disclose it
within kind of 24 hours of ithappening, but had just got all
(28:29):
our ducks in a row, goteverything buttoned up, made
sure everything was workingagain, then perhaps wouldn't
have been quite so extremelyfocused on us for a solid week
of the press cycle.
Àlex Rodríguez Bacardit (28:42):
One thorny thing about running
companies that are focused onemployee well-being can work
like Valence and totally goingagainst the hustle and the grind
is that our employees are notsupposed to work on the weekend
or long shifts like in othercompanies.
Right, we're not freakingMcKinsey, we're not Deloitte,
(29:03):
and our employees take their, ofcourse, the natural time off of
the weekends.
And then some employees of ours, at 2 pm, they're done because
they start at 5 am in themorning, right.
But when it comes to working ona weekend for an emergency like
that, what kind of policies doyou have in place to compensate
for that?
How do you rally the troops?
John O'Nolan (29:24):
We have an on-call team and incident management
process.
So essentially, if there's anydowntime or critical emergency,
we have a way of notifying teammembers and for our on-call
engineers infrastructure teamthat's a part of their general
job description.
There's a rota that we havesystems for managing it.
And for other team members, wejust ask people to help when we
(29:45):
need it.
And then on the flip side, onthe kind of work-life balance
side, we actually does afour-day work week and we have
unlimited time off policy.
So we kind of ask everyone tohelp out when we need it and we
also try and give people freedomto help out when we need it and
we also try and give people,you know, freedom and time off
when they need it.
Àlex Rodríguez Bacardit (30:03):
That's amazing.
Have you encountered anyproblem with people?
Because we did have unlimitedtime off in the beginning, but
we found out that people wereactually not taking holidays.
They were not taking days offbecause the founder team we were
not doing it.
So we had to start being reallyresilient and vocal on working
only 40 hours per week even thefounders right Because otherwise
(30:25):
people were not taking holidays.
Did you have an issue with that?
Or have you set a minimum ofdays a year that you have to
take off?
John O'Nolan (30:33):
Yeah, we set a minimum.
We found the same thing.
We set a minimum of two weeksper year that have to be taken
off, and if people haven't takenthat off in December, then we
closed the office at the end ofDecember for the Christmas New
Year period and if peoplehaven't taken it off by then, we
kick them out early.
Effectively, I think we've onlyhad to do that maybe once,
(30:56):
because if you say that thenpeople are more likely to then
just go and book it when it'sconvenient for them.
But yeah, I think, I think it'salways a bit of a thing as
unlimited time off is great, butthen if you can take unlimited
time off, then you you alwaysdefer it to later.
So it was part of themotivation for switching to a
four-day week was to recognizethe kind of intensity of work I
(31:20):
think, especially just inquote-unquote this day and age
being remote, crossing lots oftime zones, having customers
everywhere, having incidentsthat are sometimes out of hours
that people have to deal withand trying to combat that with
more of an intentional,conscious break for people every
single week as opposed to onlywhen they happen to have time
(31:43):
off booked.
Àlex Rodríguez Bacardit (31:45):
One.
I've been trying to do thefour-day work week and it ends
up on.
You know you have meetingsscattered all over the week.
I could concentrate in threedays or four only, and then the
fifth one I would dedicate itonly to admin work in my house,
because everything is soaccumulated and then you take
this day off but it ends up.
You're working on somethingelse.
(32:06):
But is there any otherdisadvantage you've found so far
with the four-day work week?
Because this is an experiment Ithink we ought to do eventually
, being a services company, it'smuch more complicated because
the clients pay for full weeksand they will not understand for
that.
But at the same time I knowother agencies are charging for
full days of work consisting ofsix hours only instead of eight.
(32:29):
So at the end of the day itkind of like evens out.
I don't know, but probably youcan.
You split the team in two andsome of them work one half of
the week, the other team worksthe other half of the week, and
stuff like that.
I don't know what disadvantageshave you found with this new
approach?
John O'Nolan (32:48):
Yeah, I think what you mentioned about having
other work and admin to do isone of the reasons why the
four-day week is valuable.
I think everyone has someversion of of that, whether it's
, you know, appointments orshopping or kids events or just
life maintenance needing to go,and I don't know, do a tax
(33:08):
return, talk to an accountant,figure out something with your
bank account, and then a lot ofthose things can't be done on
weekends because the place thatyou need to do the thing is
closed or you just have run outof hours in the day to get to
the, the life admin that isnecessary of all of us and is
difficult to do alongside afairly intense workflow job.
(33:29):
So we use the.
The four-day week, I think,helps give people space to use
that time where they mightotherwise feel like they're
losing the weekend, which whichreally sucks.
I think that's a big advantageof it.
The most significant downsidethat I think we're still trying
to figure out is the context gapbetween Thursday and Monday is
(33:52):
much larger than you wouldexpect compared to the context
gap between Friday and Monday,if that makes sense.
So having the three days offand returning to whatever it was
you were doing last week isslower to get up to speed and
conversations that start on aThursday difficult to remember
what we're talking about againwhen you get to Monday morning.
Or if something on a Thursdayafternoon is not quite done,
(34:15):
it's very easy to kick it toMonday, but then suddenly you're
at quite a big disadvantageafter that big gap of time has
passed.
So I think figuring out how tooperate efficiently in the four
days that are available for workbecomes much more important.
I don't think we're super goodat that yet, or we haven't
figured it out got the playbookdown just yet but it's one we've
(34:36):
noticed and talked about bookdown just yet, but it's one
we've noticed and talked about.
Àlex Rodríguez Bacardit (34:45):
Also, there is not enough data out
there explaining correlationbetween four-day work weeks and
taking less sick days off orneeding less vacations, and
maybe people who used to takethree weeks off in the summer
break now they take only two.
There's not enough volume ofdata or not enough companies
that have implemented it foryears, so the data set is not
yet relevant, but I assume therewill be something there.
I can think that if I amworking on a four-day work week
(35:09):
for a year, I will not beburning out so easily and maybe
I don't need to take those longbreaks over the summer and stuff
like that.
But I don't know.
That's my observation.
I wanted to circle back to thetransparency topic.
You've done a remarkable job atthe sharing.
(35:30):
The journey goes, the ups anddowns and all the hard things
about running this kind ofbusiness all the hard things
about running this kind ofbusiness.
In the same vein, joel fromBuffer made something really
extraordinary for me a fewmonths back when he shared data
openly about the plateau theyhit and the loss of revenue they
(35:52):
hit one year ago.
Something like that, becauseusually building in public is
great, but if you notice mostpeople building in public, they
only build in public on the wayup.
When they go down, they stopbuilding in public, they stop
sharing metrics because thatprobably causes an accelerated
degrowth.
Perhaps If people see that thebusiness is going down, oh, I
(36:13):
might as well change to anotherplatform.
Right, and Joel shared that, andthat was remarkable because not
a lot of people will admithitting a plateau or their
numbers going down in public,and I think that's an exercise
of responsibility that everybodyshould do.
I understand that publiclytraded companies cannot do that
(36:34):
I mean they have to do that butthey cannot communicate certain
things too early in order not toalter the value of the stock
and alarm the investors andstuff like that.
But more companies that arefinancially independent should
be more open about this kind ofdata.
I found myself that when Ishared some stuff not working or
(36:57):
some problems, we hadinternally some fuck-ups.
I'm very public about ourfuck-ups as a company and not
only we get a lot of engagementin social media.
It's not a metric I really carefor, but we get recognition and
perhaps better contacts.
We're contacted by people like,oh look, you shared that, that
(37:17):
was a fuck up.
It was great.
I learned from that.
I want to work with you guys.
I don't know if you notice likea spike in talent or kind of
like signups every time youshare your data.
John O'Nolan (37:31):
Yeah, particularly the.
You know it's the tried andtrue indie hacker thing.
Quite easy to move the needleon the startup mri chart in the
early days.
You know you, you do one bignew thing and mri doubles or
(38:02):
there's like a 50 change.
I think a lot of people stopsharing revenue, not necessarily
when they plateau, but whenthey just get above a certain
size.
When you pass around a milliondollars a year in revenue,
sharing revenue metrics I thinkstarts to be more of a
disadvantage than an advantage.
At that point.
(38:22):
You've been around forpresumably long enough that
people kind of already know whoyou are within the startup space
and you're not small enoughanymore to be considered an
underdog or like someone to berooting for because they're just
getting going.
But you are starting to becomea target for competitors and
(38:42):
competitors.
Having insight into your dataand your growth rate and what's
working, what's not working asintelligence for what they
should do and how they cancompete with you becomes an
increasing risk.
So I can understand why a lotof startups stop sharing revenue
as they get to a certain size.
At the same time, it's verydifficult to influence those
metrics in a meaningful way.
That tells a good story as youget bigger.
(39:04):
Any changes I make for the nextsix months, I'll be lucky if it
moves the needle by 10%, whichis just not nearly as much of an
interesting story.
But in the context of Ghost, weare very committed to the
transparency piece, in partbecause it aligns so much with
the ethos of the open sourceprojects as well as the
nonprofit foundation.
(39:24):
I think transparency goes handin hand with both of those.
But I do think you have to bemindful of the way in which you
do it and the reasons for whyyou're doing it.
For us, the reasons are thatit's aligned with the project
and the structure.
For other people, the reasonsthey do it are for marketing and
because they want to getattention.
But I don't necessarily thinkeveryone should share metrics
(39:48):
always and it's always good.
I think you should have a goalin mind for why you're doing it
and then measure against it.
But to your point, undoubtedlyit's something that we hear is a
strong signal when it comes tohiring team members, something
that they like, something thatthey appreciate, like the
transparency and the rest of thestructure of Ghost is a huge
pull for us being able toattract talent to work here and
(40:12):
it's a big pull for customerswho have, at various points, had
the rug pulled out from underthem from other tech companies
that start up, shut down.
They have no idea if they'redoing well or poorly, or if
they're running out of money orneed to raise more.
In them being able to trustthat we're going to stick around
, we're doing okay and they caninvest their time and energy
into building on top of Ghostwithout the fear that it's going
(40:35):
to sell to Yahoo next year.
Àlex Rodríguez Bacardit (40:37):
Yeah, exactly, I mean on the building
in public for indie hackers.
There's a lot of upside,there's literally no downside,
to saying, oh, we're doing 2000euros in MRR in our first month
or year, something like that.
There's no downside to that.
I have never set the limit fromwhich it is dangerous or
(40:59):
calculated risk.
One million in AR makes sense.
For another reason that doesn'tapply to you guys, which is you
will never sell your company.
But when you're sharing yourdata publicly, you're also
sharing your valuation, yourpotential sale price, right.
So that's some kind of numbersthey can use to negotiate
(41:19):
against your asking price, andso it doesn't apply to you.
But you have to take it intoconsideration if you have
another kind of companystructure.
The other question I have foryou with regards to
communication and style becausethis is one of my personal
struggles is by having thistransparency core value in the
(41:40):
company.
Where do you set the line intransparency right?
Over-communication versusunder-communication?
What's the right timing tocommunicate?
Oh, look, guys, in three monthswe'll have no projects for
everybody.
You better start looking foranother job or you know you're
going to overcome that situation.
So you just don't sharebeforehand, so that you don't
create an internal sense ofalarm and an emergency in the
(42:04):
company, right how to set aprotocol for that and a standard
in communication, especiallynot of bad things, but of
preparing people for potentialbad things that might happen.
You know, when there was COVIDhit, potentially we could lose
(42:24):
all our clients Didn't happen,but it could have happened.
Other companies shut down.
In the last 24 months, mostboutique agencies like ours have
either shut down or sold toother companies at a discount
because they were doing verypoorly, and so our employees
shared the concern like, hey,everybody is doing poorly, how
are we doing?
We're doing great.
(42:44):
But I have to let you know thatthis is happening out there.
It's very cold out there,salaries are going down, there's
a lot of layoff season andstuff like that.
But I don't know Like I don'tknow if you've got the recipe
for that or, because you lookvery calm and very you know very
thoughtful about everything yousay, it looks like it's pretty
(43:05):
calculated every time you shareit, but I don't know if you also
internally struggle with this.
John O'Nolan (43:12):
Yeah, I guess I kind of think about this in the
context of we try and share asmuch data with the team as
possible, particularly data thatwill help them be able to do
their jobs more easily.
So on day one when someonejoins ghost, they get full,
unlimited access to revenueanalytics that they can dig
(43:33):
through the entire thing, everysingle metric that we have
performance over the last 10years, and see how the company
has performed.
And it serves two functions.
One, it kind of sets everyone'smind at ease about how we're
doing, like, how are thingsgoing?
Are we flatlining, are we goingdown, are we going up?
You know, you just take awaythat whole category of questions
Because in reality, if it'stough, if your team are worried
(43:55):
about that, if they're thinkingabout that, then it's going to
be difficult for them to beproductive on what you hope it
is that they're working on.
So you potentially could bewasting energy in that area when
you don't need to.
Another thing is they can seethe effects of their work.
You know, if they launchsomething or work on something,
it can see like did the numbersgo up, did they stay the same?
(44:16):
Or if they have a theory abouta segment of customers like we
think business customers, whoare all asking for this new
feature, are growing faster thanthe other customers and we need
to devote more resources tothem.
If they have access to the datato go and validate that and
look at the segment of thosecustomers and how they're
performing and what they'redoing, then they can make better
decisions in productdevelopment.
So we try and open up as manythings as we can unless there's
(44:40):
a really clear reason not to.
We essentially default to.
Everyone can see everything,unless there's a particular
reason of some kind legalcompliance or otherwise that
they shouldn't be able to.
And then when it comes tosignaling you know, signaling
bad news or talking aboutpotential disadvantages that are
coming up we do our best as the.
(45:01):
When the team was really small,we did, you know, a weekly all
hands and we just talked openlyabout everything with everyone.
Now we have, you know, amultitude of teams working on
different things, differentpriorities.
You a weekly cadence wouldprobably be more distracting
than helpful.
So we do a big all-hands everyquarter now, sorry, a big
(45:22):
all-hands every month.
And then that includes astrategic kind of planning
session and bigger picture viewevery quarter and that's the
place where we bring up anythings that, as a company, we
feel like we're strugglingstruggling with risks that are
potentially on the horizon andthen open it up for anyone who
has feedback or questions totalk about it.
And and then we do the samething in person once a year on
(45:43):
our team meetups, which isalways a good venue for bringing
if there are difficult topicsthat need to come up around
company performance or like howwe're doing or whether we're
working well together or there'sroom for improvement or people
are frustrated, perhaps withprocesses or systems that aren't
good.
You know, the one of thebiggest downsides of remote is
(46:03):
empathy doesn't translateparticularly well through 2d
zoom calls.
Being in person makes it a loteasier to have those
conversations.
There's just much more of ahuman element to it.
So we try and always createspace for those things to come
up when we're sitting together,hanging out in person, and that
always is a good reset or a goodplace to do those types of
(46:27):
things.
Àlex Rodríguez Bacardit (46:28):
Good.
And to wrap up the topic oftransparency, I wanted to ask
because, before you share a techfuck-up that was probably
somebody else's fault, do youhave any personal fuck-up that
you've done in the company thathas resulted in some kind of
like financial?
In other words, what's beenyour most expensive fuck-up at
Ghost?
(46:48):
It has to be yours.
John O'Nolan (46:51):
Yeah, I've had plenty.
Àlex Rodríguez Bacardit (46:53):
It's a signature question of the
podcast, so yeah, I've had tons.
John O'Nolan (46:58):
It's just difficult to pick one.
I don't know like, yeah, youcould probably tie this back to
revenue somehow, but it's it'sdifficult one.
I once managed to get ourentire website banned from
google, for it might have beentwo months, oh wow, which is
pretty, pretty fucking bad.
So the what happened was wewere getting inundated with spam
(47:19):
.
So these and this enterprisingseo agency in malaysia had
decided that they were going toexploit ghost's free trial
system so you could sign up fora free trial which would give
you a website, and then theywould put loads of porn the
website and try and use that forSEO.
(47:39):
I'm not entirely sure what themotivation was, but it was.
Maybe it was an SEO agencyworking for a porn company.
It's not entirely clear.
In any event, we were like afive person company.
I think this was 2014 or 15.
And the resources that they wereusing up with this spam, this
automated spam, was intense,like it was costing money.
It was making servers go downbecause of the level to which
(48:02):
they were doing it, and so,eventually, how did I do it?
I found through some sourcecode.
I can't remember.
I found the agency in malaysia.
I called them up 2 am in themorning.
That time I clearly worked somedude up, presumably the owner
of the agency, and I said please, can you stop spamming us?
We, we need this to end.
(48:24):
And I'm trying to remember thesequence of events here.
I think he sort of said hewould and then he didn't.
In any event, we had a biginflux of requests, spam
requests, and again one day, andwe were going through the
server logs, I was lookingthrough Cloudflare and there was
a big influx from this IP.
That was Google, you know, likeGoogle Cloud Platform, so it
(48:44):
blocks that whole IP range likeno more VPS spamming our servers
for you.
Anyways, long story short, thatwasn't Google Cloud Platform,
it was Google Web Crawler tryingto index our site who had just
been banned from absolutelyeverything.
It wasn't until, I don't know,a month later or something, when
I Googled something in our docs, that nothing came up.
(49:04):
That's strange, it always comesup.
And then I realized that noneof ghostorg was coming up, and
then I realized that we were nolonger ranking for a single
thing whatsoever.
And that was entirely on me anda hasty click in a Cloudflare
dashboard.
Fortunately it recovered justfine, great.
Àlex Rodríguez Bacardit (49:23):
Well, thank you for doing this,
because actually we need moreexamples of that.
People think that othercompanies work perfectly and
seamlessly, while our personalprojects are always like shit,
but I think we make a prettygood job at democratizing
failure in this sense.
Look, I was worried because Ihad another block of topics that
was how to more like productstrategy and stuff.
(49:44):
We will not be covering thisbecause we're running out of
time, and one of the keyquestions there was, like I
haven't found the formula tosort of resell Ghost as an
agency, because for me it lookslike a very good platform to
build projects on.
But there are certain thingsthat you guys are so opinionated
on like the multi-languagesupport, natively or not adding
(50:04):
content blocks and stuff likethat that other platforms use it
, and you guys have got verylegit reasons not to give them
out of the box, and so I wasgoing to go into that.
But I think from this answeryou've just given me, I've
inferred how not to use Ghost,which I think it's as valuable,
if not more so.
Well, let's just wrap it here,because you've been extremely
(50:26):
generous with your time.
I'm rolling out the carpet foryou.
You got one minute left to sayhow we can help you, how we can
help ghost and any announcementsor anything you want to share
with our community love the hotones reference yeah, if you want
to check out to get it?
yes, I would.
John O'Nolan (50:47):
if you publish online and you have an email
list or you have thoughts toshare with the world, check out
ghostorg.
We're doing a lot ofinteresting things.
The most interesting of whichthat is coming up, that is kind
of on the horizon, is activitypub-based federation.
So soon all ghost sites in theworld will be networked with one
another through decentralizedprotocols that will give you a
(51:09):
social graph inside anindependent website, so you'll
be able to like reply andinteract with other publishers
and audience out there in thebig wide social web, which is
very exciting.
We've been working on this inbeta for a few months and it
completely changes the shape ofthe product in a way that I
think in the next year or two isgoing to be pretty
(51:32):
transformational.
So that's one of the fun thingswe're working on and, yeah,
love it.
If anyone wants to check outthe platform, shoot me a message
on any of the social platformsat Jono Nolan If you have
questions or want to chat.
Àlex Rodríguez Bacardit (51:44):
I totally endorse that as a
personal user and before you go,you got it coming.
Favorite Hot Ones episode sofar.
John O'Nolan (51:51):
I mean, the DJ Khaled's always a classic right
so far.
I mean the dj khaled's always aclassic right, just for for its
incredible failure.
What was the one I watched justlast week?
Oh, ryan reynolds, that one wasreally good yeah, that's,
that's very good there's a few.
I'm struggling to think of themall on the spot.
Jimmy fallon is the one I justwatched.
He was very funny.
(52:12):
What's yours?
Àlex Rodríguez Bacard (52:14):
Shaquille O'Neal, probably Like I don't
understand most of the timewe're speaking, but the faces
and the entire performance issuperb.
That for me is like an all-timegreat.
John O'Nolan (52:25):
Ricky Gervais.
The Gordon Ramsay one was goodas well, just because of.
Gordon Ramsay because of howangry he got throughout the
whole thing.
Àlex Rodríguez Bacardit (52:35):
Like why angry you got throughout the
whole thing, like why do you dothis exactly?
Exactly they're.
They're really good ones.
And then I'm really surprisedby people who you wouldn't
expect it but they toleratespiciness so well, like charlize
their own, for instance and I'msurprised gary venetro.
Yeah yeah, exactly, some peopleare like yeah, yeah, spicy, but
you know, I'm like yeah, holyfuck like I.
I have the bomb, the bomb, Ihave it in my fridge.
(52:56):
It's like literally unbearablestuff like that I love spicy,
and that's why I buy all ofthese terrible sauces.
And when I see some of theguests that they don't break a
sweat, I'm like, no, I don'tknow how they can do it.
I don't know how they can do it, but anyways, I don't know how
they can do it.
I don't know how they can do it, but anyways.
John O'Nolan (53:16):
I'm a big Hot.
Àlex Rodríguez Bacardit (53:16):
Sauce fan.
Yeah, love having you on andcongratulations on what you've
built.
Much success for the future.
And also thanks for getting theHot Wins reference first First
person in 150 episodes.
So far that gets it.
John O'Nolan (53:33):
Amazing, I'll take that, thank you.
Àlex Rodríguez Bacardit (53:36):
Thank you very much.
Popular Podcasts
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com