Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
None of the organizations I've worked withhave that really in place to start with.
And they they sort of poo poothe idea completely when I do raise it.
So I just let it go.
And then we work on the other stuffand at some point they are now saying,
Oh, do you know what we need?
(00:20):
I'm like a product manager?
They're like, Yeah, product manager!
I came up with that idea,you know, is what they...
So you got to,
you know, and they get there oncethe rest of the pieces are in place.
And I'm sort of like a little bitmore relaxed about that.
I'm Jason Harmonand this is API intersection
where you'll get insightsfrom experienced API practitioners
(00:42):
to learn best practices on thingslike API design,
governance,identity, auth, versioning and more.
Welcome back to API Intersection.
I'm Jason Harmon, the CTO at Stoplight.
(01:03):
As always, your host.
And today it's yet another old friend and
virtual neighborfrom when I lived in Barcelona.
Mark Boyd. Mark, thanks for joining us.
It's great to be here.
Good to see you, Jason.
So, you know, for me,
Mark was always one of the early folksthat was writing about APIs
(01:28):
in just all different avenues,all different places.
But kind of this whole notionof the API economy
and that sort of thing,I feel like you were one of the early
sort of most prolific bloggers
on this subject,which seemed to keep getting upgraded. So.
But now you're doing big crazy things.
So tell us a little bitabout your journey with APIs
(01:49):
and what took you towhat you're doing now.
Yup, sure.
So before the world of APIs,I was actually working urban planning
and public health and designingdata systems
for local governments to be ableto measure population health outcomes.
And I had
when I moved from Australia to Barcelona,I had a job lined up
(02:09):
where I was going to continue doing thatand it fell through.
So I was staying in Barcelona.
I was like, How am I going to make money?
So I opened my laptop and was like,How do I write for the Internet?
And so I started writing to the Internet,but very quickly
fell on to API and was thinking,This is what we needed.
We're more designing
those data models for local governmentbecause if we could pull in all of this
(02:34):
data from different areas via API,we could build the dashboards
that allows the city government to knowwhen to invest in preventative measures,
when it's where theywhen they're over flooded
with fast food outletsand need to change up the environment
or when to do the walkable streets,all of that sort of stuff.
You know,
APIs would have been the solution becausenot only could you bring in the data,
(02:57):
you could map it as well to all of thesereally crazy, interesting stuff.
So from that angle,I sort of instinctively
knew the power of APIs and just startedwriting for sites like Programable Web
and the new stack to talk about APIsand was somehow able to find this balance
between some of those technicalaspects, the business value,
(03:21):
and then also and also describesome potential use cases.
So I had that sort of, you know, magic mixthat meant
that I could write really well about itsort of instantly.
And I love doing it.
So I sort of grew from there.
And because of that knowledge,then that led me to opportunities like
Program Chair for the AustinAPI Strategy Conference
(03:43):
and then going on to do doing thingslike being invited to
do some consultancy work around APIsand so on.
And so that's how I sort of likebuilt out my work that way to now doing
more sort of API specific projects
for a range of different bodies.
Yeah.
So your company now is called Platformmobile, right?
(04:04):
Yeah.
And so tell us a little. Bitabout what you do with that.
Sure.
We're a team of nine, so one of the issueswhen I moved from there
writing about APIs and then started doingconsultancy work was
I was getting a lot of different projectsbeing offered to me.
And so it was sort of it was difficultto know which ones to prioritize.
(04:24):
So I actually went to a business mentorand talked about
what sort of values didI want to have for my business.
And I really came across the understandingthat for me I was interested in systems
that allow everyone to participateand co-create their own value.
And so openecosystems is a model of using APIs
(04:45):
where if you've got those API components,whether that's access to data, access
to web services, you can actually open itup so that everyone can be able
to build what they want to buy,what they want to,
you know, customize productsand services to their own needs and so on.
So and that allows new market entrants in
rather than the established playersalways getting the market share and so on.
(05:08):
So I wanted to do something that was
taking APIs to that next level.
And so for me it was abouthow do we support the development
of open ecosystems?
And so then we wanted to work in key oneslike banking and finance, where
there was just the emergence of the openbanking and API world health.
Because of my background in health,I did want to do sustainability.
(05:31):
I tried to do sustainability too earlyand we were not ready to grow that fast.
But that will be the next one.
And then governments, because of mybackground in government, work as well.
So. So yeah.
So now we focus on
banking and finance, health and open
ecosystems generally.
Yeah.
And how do you build those?
(05:52):
How do you support that growth?
Yeah, for sure.
So looking into this a bit beforedoing our homework and chatting with you
before we started recording here,
it reallysounds like you kind of developed a method
for these sort of segments of the marketin how to build their API programs.
Yeah, it's reallywhat's what I find crazy with banking.
(06:14):
Let's talk about banking and financeas an example.
So what I found really interestingwith banking and finance was like,
say you've got Europe and UKled on this whole idea of banks.
It's too much of an oligopoly.
So we're going to force them to have APIsand then APIs will allow
new market entrants.
So you've then got all the fintechcoming in to be able to build
more than just,you know, banks haven't really
(06:38):
modernized or innovatedvery much over the past 100 years.
You've got a savings accounts, you know,
and you've got a lending productor whatever.
So by having APIs that was enabled lendingand all these fintech
to be able to either provide me servicesor other sort of products, you know,
for particular target markets know.
(06:59):
And so that that was the reason for yousort of thing.
And then you see these
particular regulations like Indonesiasays this is about financial inclusion.
For example, Brazil as well.
So there's all of that.
But when you so regulate this,we introduced all of that,
but then there's no way to actuallythey're not
they don't have a systemto actually measure.
(07:21):
How is it actually creatingthat level of consumer choice?
Is it actually increasingfinancial inclusion?
So we tried to take a step back and say,okay, how does value get
generatedand distributed in an open ecosystem?
So we looked at,
you know, like, say, regulators come in,then you've got you know, they're saying,
okay, banks and fintechhave to open APIs and from APIs.
(07:44):
Then the level of developer experience,the level of security,
the level of digital readiness will mean,
you know, if you've got strongdeveloper experience, for example,
you're going to have more API consumersbuilding products quickly.
If you've got less developer experience,you're not going to get that
flywheel effect and that sort of thing.
And then you need
(08:05):
other things like API governanceand standards to be playing the part
and so on.And so we mapped all of that out.
And then so going backto my local government data model days,
we sort of made them thought,okay, what's the indicators or what
would you measure so that a regulator,for example, would be able to say,
yes, this is open banking, is generatingthe things that we want.
(08:26):
And as we did that, we realized, oh, okay,if you're a player in that market,
if you're an API, if you're a fintechthat wants to use bank and other
fintech APIs to build your product,
you could use this mapping to see wherethe opportunities in that market are,
what gaps there are, who's being servicedand who's not being serviced,
(08:48):
and be able to sort of jumpin on those opportunities as well.
So, you know, all of thissort of ecosystem mapping was sort of
a new endeavor that I don't see thatnot done very much anyway.
And like it really is insidethe mix of ecosystem
mapping you see is sort of the CB insight,sort of
this is how much is being investedin a particular subsector of the market,
(09:12):
whereas this is much more abouthow does how do APIs flow
or generate this flow of valueand who's missing out
and where are the opportunities andwhere could we invest more in and so on.
It sounds like you're lookingsounds like you're looking more at sort
of the network effectsand how to measure how that converts
(09:33):
into sort of economic opportunity versusthe alternative.
Absolutely. Yeah.
And it's crazy that I haven't We haven't.
So, you know, we've found some modelswhich we draw on because we sort of
look through the evidence baseand we were always updating those.
And there's some really interestingresearches out there, but there's not.
But we don't see enoughof it just yet, you know.
(09:54):
And so I harp on and on.
I'm sure listeners are tired of hearme say it, but like I always try to
quickly disambiguatebetween which platform
are we talking about, the Harvardor the MIT version of the world.
So there's sort oflike the MIT for engineer types
is like you're building a component sizeddistributed system, right?
(10:14):
We all get that. It's easy to understand.
But then, like the Harvard'sother world team
and you see this more with the MBAcrowd is like we're building marketplaces
that have these sort of,you know, at least two sided
participant sets.
And that's where you tendto see more of that network effect
and all that kind of stuffget calibrated a bit,
(10:36):
I guess.
Where
where do you feel like
you're falling when you're looking at thissort of ecosystem development mapping?
I think more on that marketplace
side of the equation.
So there's that we go a little bit furthereven because we sort of
then think, you know, my backgroundalso with the health stuff
means that I'm focused on equity.
(10:56):
So we're very much looking as well as who
potentially misses out because the problemwith APIs that I don't think
is being takenseriously enough is they are
that velocity impact of APIs is huge.
So if you go off down the wrong directionor if you go off
in a particular directionthat favors some players
(11:20):
with APIs, the speed at which you can godown that path and the mass
market appeal that you go downthat path means that you could actually
seriously lock some sectors outand some participants out very quickly.
You know, so so there isthis need to constantly keep that in mind
in some of these marketplace modelsand so on.
(11:41):
Don't I don't think they've gotthat equity lens and I've seen it also.
I don't think that seen thingsthrough the indirect beneficiaries
as far as society, the environmentand the local economies as well.
So I would like things like that.
APIs that we use the transport
have an impact on where they can be usedfor route planning,
which in turn reduces carbon emissionsbecause you're able to better
(12:06):
map carbon emissions, you know, better,
better use of resourcesto reduce your carbon footprint
because you are able to get thereand you're not like circling in blocks,
you know, in the car or whatever,you know, and you see that a little bit
with thathow some private jet APIs as well.
You know, it's greatthat we've got access to that data.
I think we need to discussthat a lot more.
(12:26):
As far as the fact that there is the like,that's where the carbon
footprint of the plan is coming from,you know, things like that.
But IP so APIs do have this real valueto be able to draw attention on.
You know, where can society improve accessand opportunity for everyone?
How are they being used?
FinTechs being huge, itbeing enabling local economies to develop
(12:48):
because they've been able to buildemployment hubs for people in areas
like where you're living,you know, And so on.
And so and you know, this environmentalfootprint and some of those models
don't factor that those sorts of impactsknow, I don't think enough
you know because they give as alate stage capitalist model.
Yeah I mean we're I think
(13:09):
most of us have been somewhat subjectto the social experiment that is platforms
like Airbnb and Uber, where it's like,you know, you see entire
the entire city compositionschange as a result of these things.
And to your point, like it'sa digital access to or that many don't,
just don't have an entry point to drive.
Yeah, absolutely.
(13:31):
So it sounds like you've kind of got your
got your base going in fintech,which makes a lot of sense.
I mean, I know
from a stoplight perspective and just,
you know, my view as a whole,like the last few years has been quite
a huge investment from banking and fintechand certainly open
banking and all that.
But what I find
particularly interesting is this notionthat you've kind of come up with some
(13:56):
approach to
governancethat is generally somewhat portable,
or at least by the sounds of it,between these different industries
to include government,finance, banking, health, so on.
Yeah, Yeah.
So what's what are kind of the,
let's say,pillars of the kind of your approach?
Okay, cool.
So then we did some workwith European Commission where we looked
(14:18):
at 343 to be exact
articles in reports and
published journal pieces on APIsin the best practices and so on.
And then we whittled that down,I think to just under 200
where we that we were able to drawon, to be able to then say, okay,
this is the evidence base we've gotas far as what's worked with APIs
(14:42):
and so on, both in governmentsand in industry.
And then out of that we,we developed a framework,
the European Commission, which is the APIframework for digital governments.
But I think it appliesjust as much to business
or any sort of nonprofit organizationsand so on as well.
And that's really talkingabout sort of pillars at the core
(15:02):
of the policy tacticaland implementation levels.
And so at the policy level, you do needyou know, I know we talk a lot about that.
You know, you need the sea level
sponsorshipand you need to have that high level
to be on board that there islet's move to an API approach.
So you do sort of need that.
(15:23):
I think, but you definitely need that.
But like Wade,then I think what falls down
is when we get to that technical levelwith those line of business managers
who just have too much other priorities onand like they lost out on that,
they may be sold on that in the sensethat,
yes, they knowit's an organizational priority.
And yes,
(15:43):
we do believe in that IP.
I want to get that.
They want to get their bonus, too,and follow the rules.
Yeah, right. Yeah.
You know, but the fact the fact is they've
also got all of these other prioritiesthat they don't immediately see.
APIs could be the ones that helpsolve them, solve those problems as well.
And when they do, they're like, Oh,but it's so much work to just re-orient.
(16:04):
I did some work with the bankis it years ago?
And they were looking at they had their leasing project, they had sort of a leasing
line of business and they thoughtthat would be a good one to start with.
IPOs knowsaid they sort of wanted to create an API
and we gave them a ton of use cases,you know, like you could help
move people onto electrical vehicleslates, you know, sort of thing.
(16:26):
For some of your corporate customers,you can be better cycling through cars
that haven't been used once or twicea year or so, all of this sort of stuff.
So they had all of thatand they were like, Yeah, let's do it.
Okay. First of all,you've got to have your data set.
And that had a traditional data setthat they were going to plug into the API
and then they realized I had to clean upthat data set to be able to feed it.
(16:50):
And I were like, Oh,
so that
pretty much just killedthe whole very game because when they sold
the work that they had to do aroundcleaning up their data sets in order
to have something of valueto pass through the API,
it was it's like when you get at home,people
don't want to join a businesswhere they're going to be
(17:13):
dealing with legacy code,where there's not good documentation
that will refuse a job, will leave a job,you know,
if they've got to disentangle someone'slegacy code and it's sort of like that.
So those middle managers are constantlymaking
that decision of like,do I have to invest in that now?
Can I?
Because that's going to be six monthsto just clean that out before
(17:35):
we start to get any business valueout of the API.
And so up until recently, I think
it was very easy to decidethat's too long a time frame.
So they will do ideas on the next project,but not on this one, you know.
So there was so thatthere was that tactical level.
And then there's the implementation level,which I feel like is almost the easiest.
(17:57):
We know pretty clearly whatthat API design
based practices are and all of that.
So even security issues, we we
know that there'sa need for particular tools in that.
But it's a, it's a known problem
know you knowthere are nine solutions for.
Engineers have been building APIsfor ten years in this the
the amount of tools availablefor engineers now to build the API.
(18:20):
So like I was like don't thinkwe need tutorials for that anymore.
There's so much out there I agree with youon implementations the easy bit.
So then that so then that means sowhen we then work with businesses around
and multinational
multilateral organizations,which is a lot of our business as well,
then we're trying to figure outhow do you create API governance systems
(18:46):
that aren't too bureaucratic And the done
also require you to refactoreverything you've got already as well.
So often it means that we've got to takea two track approach
where the APIs that they've got, you know,which might be like 30 to 50 external
facing APIs, for example, something,you know, that's not unusual number
(19:09):
that were each built individuallyand differently and some use three digit,
three letter country codesand now these two used the two letter.
It's all of that sort of stuff.
We've just got to leave thosebecause the the business system
that we're integrating with them,they don't have the investment to upgrade
their IP, you know, to re integratea new version of the API or something.
(19:34):
So they've got to sort of likejust accept that
that's it and then sort of go, okay,but we live in a new world
now and going forward,this is how we're going to build APIs.
You know, it's.
You know, I feel like thatthat particular little bit there of like
except that what you have ismaybe not that great, but it does
(19:54):
the job and people are using it,move on and build something new to like
that's the first humpto get over everywhere I've ever been.
And the idea that you're going to go backand fix everything you did
before is just like it'sprobably not going to happen. So,
I mean, that's like
the thing that people just need to get getokay with fast, right?
Like you're going to definea new archeological layer
(20:15):
in the history and, you know, go for it.
It's funnybecause we use terms like technical debt
and someone say,I heard someone say once legacy
or in other words, things that are workingwell, you know, like so we see that like
the thing is that they areif they working you know sort of thing.
So we've got to sort oflike not have that.
(20:36):
Tim taken off like let's rebuildeverything all over again, you know,
sort of thing.
Now that we, you know, this new way,you know, and,
and I think giving to me
when we've given businessesthat permission to keep what they have,
then that's been easierfor them to be inspired
(20:56):
about getting on to this new visionfor the new stuff coming up.
You know, and that's where we've seena little bit of that shift
on things like data governance.
So, you know, you've got API governance,you know, trying to get
API said that
they were all build
to a consistent standardand that they might be use the same
data models and that you knowthey've got the same nomenclature
(21:17):
and that they use in the sense of designprinciples and all of that sort of stuff.
But also
then you've got this data governance workwhich we when we've looked at businesses
and organizations, they often saythat is completely separate.
So we've I've worked with companieswhere we're doing API governance
and they're like,Oh yeah, data governance that's being done
in a completely separate project.
(21:38):
And yep,you've got data models in your iPods
and they're like, Yeah, well let's justkeep what we're doing sort of thing.
So we're trying to bridge those twoworlds as well a little bit.
And so yes, so then you've
got to work with themaround the data governance side of things.
And they areI mean, I love some of the stuff
that's come out of governmentfrom UK around civil services
(22:02):
has been really inspiring enoughto on that quite a lot.
So things like, you know,like what basically used to happen
or still happensin a lot of places in governments is H
Everyone wants to work with schools,
you know,so you've got the Education Department.
Sure, they working with schools.
But then, you know, the EnvironmentDepartment wants to describe
a new recycling program and someone says,let's let's deliver it through schools.
(22:26):
You know what then is the nursing program?
It's like, let's deliver thatthrough schools or whatever.
So then all of these other departmentscreate their own schools list,
you know, and their own databasefor schools rather than using the MH Kasam
database of the schools, you know,because they want different
fields and, you know,they don't need the principals
knowing they need this other teachersname or whatever, you know, sort of thing.
(22:50):
So you've suddenly they've got like 20
schools, databasesacross various departments,
some of which only one of whichis probably maintained well.
And that's one in the Educationand the Education Department's public
of multiple out of out of sync databasesas well.
So, you know, like the first thingand this is from that UK digital services
(23:12):
is you find who should be the digitalwho should be the data registrar
and they're the ones whothen are responsible for maintaining that.
And then all of the other departmentsand everyone else then uses that data
set via APIso that it's, it's not like you.
We've made a good bulkdownload of that data set.
(23:33):
It's more than that. It'sno, it's real time.
You're just using that one, you know,so that we moving to,
moving themtowards that around the actual data sets.
But often it's the data models as well.
And just having like a set of might be 20.
And this goes back to the
idea of golden paths and the
and that sort of page roadsort of idea is like, okay,
(23:55):
what are the ten or 15 most commondata models that most APIs use?
Like if it's a country card,
are we going to do toto let the country go or straight.
You know, and so you decide on thatand then you've sort of build up
gradually to have those sortsof data models as well, you know?
Yeah, establishingsources of truth, right.
(24:16):
And like it's I feel like that'sa pervasive issue is like,
you know, that nobody's really clear
on who owns the dataand so everybody's got their little horde.
So yeah, just establishingsources of truth absolutely makes sense.
Yeah.
I don't know how much it worksin business.
Maybe a little bit,but definitely in government.
The issue is also that you don'tif you've been spending money
(24:40):
on developing that data setand now all of a sudden
you're going to use a common data set,You don't want to give back that money.
So that'swhere it sort of stops at the moment.
I always think everyone's like,No, we want to keep our data set because
you can't use that then as a way to saywe don't need all of our budget.
(25:04):
So there needs to be a gainwith that legacy and a new world order.
We need to sort of accept that, okay,you can't maybe you get maybe
there's incentives where you get to keepyour budget advantage.
If you show you're reusing components,you know, but then you see APIs
completely impactingon financial modeling for an organization.
(25:27):
So the
Yeah, and I think you hit the nailon the head with the phrasing there
that like, how do we build incentivesfor shared leverage?
I do think it's veryportable concept in business,
especially if you have a historyof a very siloed organization
where it's like,we do this thing and we have to fight
and claw for every bit of budget,you know, for giving stuff away.
(25:51):
That kind of mentality,
it causes problems.
I've been harping a lot lately that
as companies transform their platformand thinking right the way they think
about their business, that stuff one isexcept that your your development,
your engineering should be treatedlike a giant open source project.
(26:12):
You can't put walls between technology
and that certainly pervadesinto a lot of different layers.
But that's all just to confirm that.
Yes, I think that totally is a problemin private industry.
Probably a little different.
But nonetheless, that sort of siloedthinking and protecting from different
divisions and stuff can really hampersomething that, to your point,
(26:37):
could fundamentally changethe way business is done.
And in order to make that transition,everyone's
going to have to kind of pull togetherand look at things different.
So that side of things,we try to introduce the inner
source concept and that sort of mindsetinto an organization.
And as far as I'm getting at the moment,
so it my successes are huge on this.
(26:59):
Like,I don't think I've solved that just yet,
but what we're doing is just simple wikis.
So for example,anyone who's building an API
in a line of businesscan share that on a common calendar.
Or you know, I said that there is acrossthe organization, you can see in one place
all of their APIsthat are going to be built this year.
(27:21):
And so that if you're building an APIthat's got a form
element to it, you might want tolook at what other APIs are being built
by other lines of businesswho are also doing having forms
in, you know, going to
the interface is going to be a format the end, you know, sort of thing.
And so work with some of those.
So, you know, we're just trying yetto try to get that mindset
(27:42):
that you're just talking about,sort of fostered by having that.
Another one is a taxonomy.
So showing the APIsand some of the functionalities like what
in one organization, there's multiplelines of business, but a lot of their APIs
have checking the status of an applicationas part of the API.
(28:02):
But all of themhave been built differently.
But they've all got thatsame functionality.
So again, having a taxonomy is like,okay, if you're building an API
that's going to have checking the statusof an application, who else has done that?
Which one is do you think is the best?
And again, that might be that we moveto a data model or best practices for.
So eventually they've got a datamodel of here's how you do
(28:25):
a status application check, data model,you know, sort of thing.
But they might not have that at first.
You might just have a look at the fontthat they've already got and decide which
ones you like, you know,
And that's what we
mean by the lowbureaucratic approach as well.
You know, you can't come in straight awayand say like because otherwise it's like,
let's have six months of meetings todesign the data for a status application.
(28:49):
You know, I.
I left because I've been subject to that,
some of which.
Yeah.
And it's funnybecause I think I tend to have,
you know, my sort of engineeringbackground betrays me.
And, you know,I always talk about this like
even if you have some kind of managementmandate
(29:09):
to your point,it's really a kind of mid management.
And in larger places that's typicallyyou're like the senior VP,
VP ranks that they're the onesthey really dealing with.
And I don't know, may them call me like
cynical,
but at this point I'm like,build a band of rebels, of people
who get it in product development,ghost ship something
(29:30):
and make them look badand they'll come around like that.
You know, it's way too adversarial.
This is why I'm not in consulting right?
But I always appreciate and respect folks
who workwith kind of the more business side
to figure out how to turn their heads awayfrom pipeline business
(29:52):
thinking into platform thinking becauseit's a different way of doing business.
So I love it. And yeah,yeah, you're right.
And like I said,
I mean, I'm a very passionate personand I get in, you know, I love this stuff.
I do, you know, sort of thingsnow, but I've had to learn not to
fight every fight
as soon as it comes up, like in my big oneis where I'm product managers for APIs.
(30:16):
Like I come in always when we're talkingabout API governance and say, you know,
you've got to have a product manager.
And they're like, No, you know, theyyou know, none of the organizations
I've worked withhave that really in place to start with.
And they, they sort of poo poothe idea completely whenever I raise it.
So I just let it go.
And then we work on the other stuffand at some point they are now saying,
(30:40):
Oh, do you know what we need?
I'm like a product manager.
They're like, Yeah, yeah.
I came up with that idea, you know,I like his website.
So you've got to,
you know, and they get there oncethe rest of the pieces are in place.
So I'm sort of like a little bitmore relaxed about that.
I think one of the biggest principlesthat working on big platform projects like
(31:02):
this taught me is like walk in the roomknowing what you're willing to lose.
Like you can't you can't win every battle.
You just need one thing to shipthat's really, really good.
Yeah, if you can get to that point,the rest will turn for you.
And then I A side note too, on the APIproduct manager thing.
Yeah, we follow that this topicpretty closely, especially on titles,
(31:25):
just because titles have always beenmeaningless in platform work. But
API Product Manager on the Risethe last couple of years for sure.
Like well, we're seeing more of it
and I almost never saw anyonewith that title before.
And now we see it all the time.
We even have stories of,
of product managers who design
the APIs without engineering involvementfor the first draft.
(31:48):
So I think in the private sector,we're really starting to turn a corner.
But good to hear that there's at leastglimmers of hope in the public sector.
Yeah, yeah, yeah, absolutely.
So let's see, we touched ona couple of different things,
I think around kind of business value
and marketplace equityand all these sorts of things
(32:11):
as well.
Some of the cultural change on innersource and some of this stuff, I guess.
So there are other, you know,and you sort of referred
to the fact that like the implementation,it's easy, but I guess
are there any other sort of facets of thisthat you think are essential?
Um, so the
so I mean, with going back to that pageRoads idea, we're trying to do the whole
(32:32):
internal developer portal, We help buildthose as part of that API governance.
They're somewhat, I would say somewhatshitty thing about that is that gotten us
really seems to be really gungho on back stage
as a internal developer portal and add
you know, it's very good at convincingeveryone that they should be using back
(32:54):
stage and it's really horrible to use,
you know, where, where we often build,
you know, we're building these dice
internal developer portalsfor using backstage because that's what
organizations want, you know, Butit's sort of we're sort of stuck with it.
But, but it's sort of like,
(33:14):
I guess
so I mean, there is that needfor like an internal catalog
or internal inventory of likewhat do we have?
And also when we start a project,how do I just get my project set up
and ready to be able to start doing stuffrather than, you know,
getting the environment rightand all of those so getting, you know,
getting even just little things like,you know, for us with one client
(33:38):
that's aboutif they're going to be working on
the new Approach project,then making it easy for them
to be able to alert that they should beone of the user seats in the stoplight,
you know, subscription,those sorts of things that can
if you don't have it all in one place,it might take a couple of weeks
to figure out thatyou made all of those things, you know,
So we're trying to sort of bundlethose together,
(34:01):
which is something we've learnedfrom team Topologies
and they are approaches to sort of,you know, making,
you know, and the like the Pipedrive stufffrom like Netflix and
and so on around that sort of golden typeroads sort of ideas.
So yeah, so that will be another one,the internal developer portal
that's working on the standards,
(34:22):
whether that's the API standardsor the data models.
So trying to bring that together.
And then what we say,and this is what I love about Spectral
is that you can then put those,
those standards into spectralso that you're taking out.
So we don't in a lot of places,you know, the governance committees
(34:42):
I'm big on, you know,
I think they should be structuredthe way you've got a governance committee
or something like that.
But that's just normallythat's like the architectural decision
task force or whatever
they can approve.
You know, they should be improvingthe API standards if you like,
but then they don't need to approve a APIthat gets built
(35:03):
because if you've automatedthose standards, then they're influencing
that decision without it, without thembeing a bottleneck for them having to
then go through.
When you wait for the three
waits for them to meet and all of thatsort of stuff to get it past all of that.
So yeah, so, you know, like we're tryingto look at automated systems like that,
(35:24):
The inner sourceto be able to said that everyone's
sort of more willing to share informationabout what they doing,
writing little notes about why
they've made decisions,particular data model decisions and so on.
When there is an a clear data governance
decision made
to get higher up the chain,those sorts of things.
(35:45):
Yeah.
And then and allowing that sort of legacyversus new newbuild approach
to sit side by side, you know,I think that two together they have
the sorts of approachesand working from that tactical level
is really important.
They're cool.
Well,
I don't know how much you're still writingor publishing things
(36:06):
or other places people can go to kind offollow what you're working on.
Sure.
There is platform on vogue.com,
so we've got a blog therewhere we talk about a lot of our models.
So we've got stuff around our openbanking model.
We've done some work with World Health
Organization with mapped out and openhealth model as well.
That was now worked on.
I'm releasing some stuff on data,outdated data governance systems,
(36:28):
and I've got a blog coming out shortlyabout how
the API governance model as well,so that a lot of that.
So we do sort of thatsort of thinking work
and try to sharethat we're back to rate you
so anyone can come alongand say at any time our website
we're going to do a complete refreshwhere it's
moving more towards a resource hubfor open ecosystems as well.
(36:51):
So you can be checking that outby the time this is going to air.
Probably 13.
Well, I guess I'm going to come backand this is always,
I feel like a quiz as to whether or notyou've actually listen to podcasts before
because you either see it comingand you have a good answer.
I'm going gonna catch you outand I love to catch people out. So
(37:12):
we talked about all this stuff.
I mean, it's a whole lot of basesto cover to really get a platform
scaled out and governedand designing at scale and all that.
But for listeners who are going,you know, we're just getting started,
we want to kind ofhow do we get the ball rolling?
Like if you had to startif you go into a client from scratch,
(37:35):
what's your go to Starting Pointif you don't know anything else?
Sure.
Then I think it's about finding
where who's got ownership
and who's got the interestand working with them first.
So there will beand so what might be so generally
this has been in one of two places.
(37:56):
It's either the
the the person who's hiring usor has got the remit for being able
to spend a budget and they want to be ableto get the wins on the board.
So them but then quite oftenthey're lost with them way
too way to get it up and that there's,there'll be someone in a line of business
(38:16):
that or a team in that line of businessthat's building a new IP or doing
something that can be our pilot projectso we can demonstrate
how when all of this works together,it works really well.
So we try to find that projectwithin an organization
and then work with thembecause they're often quite excited.
They are often quite new and they arewilling to admit that they don't know.
(38:40):
So in some stuff.
And so they're eager for the extra help
and that and they're wantingto also leapfrog and use
some of the best practicesor the newest tools or whatever as well.
So they're super arcane.
So we try to
sort of find that team, and that's the waywe've sort of had success in both.
So we're then working in two levels.
We're working with that product.
Oh, you know, the ANA, the client, Ana,who's sort of hard to see and
(39:03):
and then with that individual project.
So we're doing the organizational stuffwith them, which is a lot more slow moving
and then showing how that workswith that team in place as well.
So you don't think that you should godesign all the future potential APIs
and define all the standards for a yearor two before you launch something?
(39:24):
I've worked with getting some myselfwith consultants.
Who'sactually suggested doing all of that?
Yeah, but now, I mean, I'm very mucha work in progress kind of person.
So like, you know,
when I mention that taxonomy,you know, we sort of just did that.
It's not quite back of a napkin, but,you know, it's sort of
it's really rough deck work,you know, sort of thing.
And then, you know,
(39:45):
we can do the that's the MVP, you know,and then we can build from there.
But that's they didn'thave that before, you know.
So it's sort of that kind of yeah.
Yeah.
I think you have to have
a loose, big vision of where you're goingand how it's all going to fit together,
you know, I like things like businesscapability modeling.
Some people use TDD,but we did a little religious for me,
(40:07):
but I like what you're describing is like
this. We see this a lot.
Like you have the management mandateseasy to get these days,
but they don't really knowwhat they're asking for.
In a lot of cases, they're like,I read that we should do APIs
and we're going to be faster,make more money, so let's do that.
But show me how.
Right and to your point is like,go find the launch partner.
(40:28):
Where can you piggyback the roadmapwith something that already has good
momentum, really make it super sexyso that when it goes out, it's
a commercial success and technical successfrom developer experience standpoint.
And you know, in my view, communitybuilding is it's a cool game, right?
Like everyone wants to be cool,especially in the engineering world.
(40:49):
We're not cool, so we just want to emulatepeople doing cool stuff.
So if somebody did something cool,I want to do it too, right?
So want something cool andthen everybody's going to want to copy it.
And that's the best way to sort ofbuild the movement.
So I love it.
Well, Mark,
thanks again for taking the timeto join us and share all your wisdom
(41:10):
and we'll definitely written you on andand watching your blog and all that stuff.
So thank you. It'sgreat to be here with you.
Yeah, great conversation.
Thanks.
Thanks for listening.
If you have a question you want to ask,look in the description of whichever
platform you're viewing or listening onand there should be a link there
so you can go submit a questionand we'll do our best to find out
(41:32):
the right answer for you.
API Intersection Podcast listeners
are invited to sign up for Stoplightand save up to $650.
Is the code intersectiontend to get 10% off a new subscription
to stoplight platform starter or pro.
Take a look at this episode'sdescription for more details.