February 10, 2025 • 51 mins
Text us your thoughts on the episode or the show!
Unlock the secrets to building a formidable marketing operations function with insights from our expert guest, Deirdre Mahon. Deirdre, a trusted advisor in product marketing and growth, reveals the essential steps early-stage companies must take to set up their marketing operations for success. From selecting the right marketing automation system to establishing clear KPIs and fostering cross-departmental collaboration, she provides invaluable guidance to ensure your organization is on track. Deirdre also highlights the often-overlooked aspects like budgeting for tools and processes alongside campaigns, ensuring a well-rounded approach.
As organizations expand, the transition from a generalist to a specialist model becomes crucial. Explore strategies for creating specialized teams to manage different stages of the sales funnel efficiently while avoiding potential pitfalls like inefficiencies and misalignment. Deirdre emphasizes the importance of clear communication and strategy alignment to maintain seamless operations and sustainable growth. Regular performance reviews become your best ally in deciding which strategies to pursue or discard, enhancing overall efficiency.
The episode takes a transformative turn as we discuss the role of AI in modern marketing. AI tools like ChatGPT and MarkovMLcom are changing the game, automating repetitive tasks and enabling marketers to focus on strategic and creative endeavors. Discover how balancing AI capabilities with human intelligence can unlock new levels of innovation and productivity. With AI handling mundane tasks, marketers can shift from mass messaging to creating personalized, impactful campaigns. Join us in exploring these exciting possibilities and transform your marketing operations for the future.
Episode Brought to You By MO Pros
The #1 Community for Marketing Operations Professionals
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
Hello everyone, welcome to another episode of
OpsCast brought to you byMarketingOpscom, powered by the
MoPros out there.
I'm your host, michael Hartman,joined today by well, just our
guest who we'll get to Joiningme today is Deirdre Mahon.
Deirdre is currently a productmarketing and growth advisor
with Superset, advising theirportfolio companies.
She is a full-stack marketingleader with hands-on experience
(00:22):
across growth startups andpublic companies.
Particularly strong atexecuting against strategy and
driving the tactics to exceedgoals.
She has been instrumental inthe creation of new market
category segments, includingobservability, real-time data,
movement, cloud consumption,analytics and big data.
And, as she said this is not memaking this up, she said this
(00:42):
she's probably been acquiredfive times.
So, deidre, thank you forjoining us today.
Speaker 2 (00:47):
Thank you for having me and lovely to be with you,
Michael.
Speaker 1 (00:50):
Yeah, yeah.
Well, we were enjoying a nicechat about rugby before this, so
we better get back to business.
So, for our listeners, we'regoing to talk a little bit about
a number of different things,but really it's going to be
centered around the idea that AIneeds human intelligence.
We'll get to there in a bit.
(01:11):
So, deirdre, during your career, you have worked mostly in
marketing and leadership rolesat early stage companies, some
larger companies, and you'veworked closely with ops
professionals to stand up thatpart of the marketing function.
So, you know, I think it wouldbe useful for our listeners,
(01:31):
before we get into the maintopic, to kind of get your
perspective as a market leaderon what you see.
As you know, what is abenchmark kind of marketing
operations function compared toothers?
What do you look for in thoseroles?
Speaker 2 (01:47):
Yes, by that question I'm assuming, mike, you're
talking about what is a goodstarting out marketing
operations look like from ahuman side and otherwise right.
Speaker 1 (02:01):
I mean, I think you could take it wherever you want.
But yeah, I'm thinking about, Ithink there's going to be
probably two main categories,right.
One is starting it from scratch, which I think some people get
to do most, don't, uh?
And if they do do it, thenthey're, you know, they're
juggling everything, yep um, andthen maybe a little bit of
(02:22):
overtime.
And then others who areprobably in sort of a role
similar to mine.
Very often, even if I wasstarting something, it was
partially in progress, or Iinherited something-series A,
and you're right, sometimes youknow somebody else and they
could be a sales or somego-to-market human has, like, at
least got HubSpot licensed, butI mean that's just the very
(03:00):
sort of minimal viable.
Speaker 2 (03:02):
So when I think of a true functioning marketing
operations practice, I thinkfirst things is the engine.
So what is your engine thatyou're actually going to execute
?
And the obvious thing is what'syour marketing automation
system?
Most people on the call todayare probably either using
(03:25):
HubSpot or a close competitor,hubspot.
We love them because they'veactually helped younger
companies get on the engineplatform as quickly as possible
and not super expensive.
So I personally do like HubSpot.
Then the second thing is OK, dowe have an agreed upon set of
(03:50):
process steps?
And you know, if you have asales team in place or like some
small team that's actuallyactively selling and engaging
customers, then that's awesome.
So you have a partner thatyou're going to work on defining
those process steps.
And one of the most importantthings to get right out of the
(04:12):
gate is what's your model?
So are you a sales led?
Are you product led?
Are you hybrid?
Right?
So your process steps should bedirectly mapping to how people
engage and buy.
And then the last third pieceand then I'll touch on the
people side of it is.
The last piece is do you knowwhat you're actually going to
(04:33):
measure and do you have anagreed upon set of KPIs or goals
?
And how are you going to divideand conquer and co-own and have
accountability for thosenumbers?
So that's where, hopefully,you're going to sit down with
your sales leader and yourperson in finance.
Based on the model, figure outwhat those numbers look like
(04:55):
okay, I was gonna.
Speaker 1 (04:57):
I was gonna ask you if you met goals and metrics for
the marketing I've seen, formarketing or for the
organization, maybe the revenueengine, if you will.
But I think you just by whatyou just said there, answered it
that it's really more foreither all marketing or all of
the marketing sales functions.
Speaker 2 (05:17):
Yeah, Just to touch on your point about you know
goals and metrics for the opsteam.
A lot of times, especially inthe early days when you're
standing up the engine a lot oftimes it's like, okay, can you
deploy and get this new tool orsystem or process in place by x?
date train and enable otherpeople on the team to be
(05:40):
comfortable and proficient withusing it.
And also you.
You know you immediately assoon as you're buying any tech
system, you're immediately onbudget topic.
So as a team, have we carvedout enough budget to do all that
system engine and trackingstuff?
And sometimes that's like, oh,it's third level of importance,
(06:06):
because oftentimes people arethinking like, how are we going
to put money into campaigns andprograms so we can get the word
out and start campaigning andyou know outbending?
And you forget oh, don't, weneed some budget for the tools
and the process steps and theonboarding.
Need some budget for the toolsand the process steps.
(06:27):
On the onboarding.
And I will say from a peopleside, when you're young and you
have a very small team andthere's maybe two people, doing
this who have other jobs as well, by the way, that you often are
like a little drowning eventhough the rest a lot of times,
sadly, the rest of theorganization is going.
You're just building processand engine.
When are you going to actuallydo some marketing Right?
(06:47):
So you really have to sort ofstrike that balance and I highly
recommend all the listeners tolike invest early in the engine
because you don't want to begoing back and fixing it later.
Get the foundation right andthen, from a people standpoint,
pull in some specialtycontractors that can, you know,
(07:07):
be a catalyst and fast forward.
Just make sure that you pickthe right people.
You know you don't want tobreak your budget, but it does
tremendously help if you have ayounger, newer team that doesn't
know all the moving parts toall the tech, that doesn't know
all the moving parts to all thetech.
Speaker 1 (07:24):
So just curious, going back to the people,
especially as starting up a newfunction, would you look for
someone who's a little more of ageneralist and maybe not an
expert in, say, a marketingautomation platform per se, but
has the skills to do it?
Or do you look for I need theperson who's an expert in just
(07:47):
use HubSpot as an example, sinceyou brought it up but the
expert in HubSpot who haspotential to be able to do the
other things and support themarketing teams and be able to
be an advisor to those teams.
Speaker 2 (08:01):
Yes, good question.
My bent is towards thegeneralist because when you're a
small team wearing lots of hats, if you hire only a specialty
person that knows only one thing, then you'll end up you know
you yourself will end up pickingup all the other pieces.
So I want to hire somebody who'sa generalist.
(08:21):
If that person has someknowledge and experience with
HubSpot out the gate and has anappetite to learn and grow,
which, if I'm honest, mostmarketing operations folks that
I've worked with in the past.
They have ambitious careers, sothat's awesome.
I have worked with folks thatare really comfortable with data
(08:43):
and really comfortable with inthe weeds.
That that's more important.
If that individual iscomfortable doing that and then
you're on a good path and easilyshow them the growth trajectory
out of that.
They're not going to be in theweeds forever, but they have to
be comfortable with that.
Yeah.
Speaker 1 (09:02):
I've rarely had the opportunity to really build a
team from the ground up and Ithink early on I would have said
I need somebody who can run theengine, call the marketing
automation platform, and be ableto fulfill needs like building
emails, getting them deployedand that kind of stuff.
And then after that I wouldprobably want someone who can
(09:25):
take some of that like split itup a little bit, but I think
I've gotten to the point whereI'm glad you brought up the
analytics piece, because I thinkI've shifted a little bit that
I think a second hire would needto.
I'd want somebody who reallycould understand data analytics
and help us get insights,because I think there's a lot of
potential leverage from thatthat you don't get if all you're
(09:46):
doing is executing on campaignsand not to say those aren't
important.
But um, at least up untilrecently and we've had some
interesting conversations withguests and vendors and stuff
like that but um, I'm just likeI just that reporting analytics
piece is such a unique.
I don't think there's a deepenough skillset in the pool of
(10:07):
talent out there, nor do I thinkit's a.
It can be just built on top ofreporting tools, right?
I think there's an effortthat's required to do it well.
So it's interesting that youbring that up to that as well.
So that's just because I dothink it's a really, really
valuable one that's underappreciated by other marketing
(10:28):
leaders I've run into yeah, 100the.
Speaker 2 (10:32):
The reality is that marketing has a ton of data,
right?
Our systems are naturallygathering it all, and that's
from like website traffic toconversions and engagement on
your HubSpot embedded forms, tothen how it downstream converts
in the funnel.
And first of all you have toknow the process and the model,
(10:55):
and then you have to startcollecting the data and start
having some benchmark orbaseline, and only then can you
figure out where gaps are orstuckness or like, without
having exposure to the data anda comfort around how to even
read and deduce from the data,then you really can't move
(11:17):
forward.
so, and that's number one andnumber two, like the marketing
team has to get reallycomfortable with that and then
you have to be able to sort oftranslate that to other parts of
the business so they understand, because if you drown them with
too much information, too muchdata, that just glays over.
Yeah.
So you have to pull back andsay, okay, here's the most
(11:37):
important things sales leader orCEO.
So getting really comfortablewith data and understanding it
is critical in my opinion.
Speaker 1 (11:47):
Yeah, well, and I like the second part of
translating or the storytellingpart of it, exactly Doing the
analytics, all right.
So you asked me the questionback at me about what kind of
organization we're talking about.
You talked a lot now about thestartup world, like do you, in
your experience with largerorganizations, like, what do you
(12:09):
see that's different in thoseearly stage kind of companies
and what they're doing from amarketing, marketing app
standpoint, and those ones thatare, say, medium, larger, maybe
even enterprise?
Speaker 2 (12:20):
Yeah.
So the smaller stage is likeyou're going to get your MVP
right, get the engine built, andthen you start to execute and
you start to measure.
And I always think of it in aone-two punch way you invest,
you execute, you measure, andthe shrinking of the time
between those two things iscritical so you can move forward
(12:42):
.
Two things is critical so youcan move forward.
When you're further along,hopefully, you know what a good
sales cycle looks like and yourgoal is to get it repeatable and
to make sure everybody's on thesame page with your
understanding of here's what arepeatable looks like.
And then you're probably goingto be like okay, how do we scale
(13:03):
up?
So, from a marketingperspective, your operations
teams are, besides all theexecution that they have to do
and focusing in on the thingsthat work and measure.
You're now figuring out how do Ido it at scale, because it's
always up into the right growthyes um, and so in case, what
(13:25):
usually happens is you actuallystop being more of a generalist
and you start being more of aspecialty expert.
So there's, like I've seendifferent ways to skin this, but
sometimes you have a dedicatedteam that is only focused on top
of funnel, and then you haveanother team that's focused on
like sort of of funnel, and thenyou have another team that's
(13:45):
focused on like sort of midfunnel.
And then maybe even somebodywho's working really closely
with product.
If you're more of a self-servePOG and you're more sort of in
the bottom of the funnel, whichis all about usage and adoption
and expansion and renewals,right, Very classic POG.
Speaker 1 (14:04):
Some customer marketing in some places would
be yeah, exactly.
Speaker 2 (14:08):
Because you're further along and you have
existing base and you have tokeep them happy and learn from
that and upsell, cross-sell andso on, references and et cetera.
So you end up being like I likethis sort of demarcation between
am I a generalist or aspecialist.
So now you're probably more ofa specialty and you're focused
on a particular part of thebusiness or maybe a particular
(14:30):
segment, or you're working withthe enterprise sales team
directly to help them, versusmaybe self-serve side of the
house.
So ultimately, at the end ofthe day, you're honing and
tuning what's working so thatany additional marketing dollars
that you're pouring in, thatyou're being really efficient
(14:52):
and spending where you know thatthere's going to be a growth at
those different buying cyclestages.
Now you know every week whenyou do your weekly business
review.
Now you know what is yourexpected goals and KPIs and how
(15:13):
are you mapping each week tothose.
So each of those dedicatedteams can really get much closer
to the problem and makerecommendations on what to do to
engage, whether it's top,middle or bottom of funnel.
Speaker 1 (15:29):
So one of the things I've experienced and I don't
know if you've seen this is Ithink this is a real challenge.
As teams get larger, you think,I think intuitively, you're
like oh, we can multiply theamount we can get done, but I've
found that that multiplier ismuch smaller than we think it is
because there's be therebecomes inefficiencies in how
(15:51):
things get done, because nowyou're, instead of you had the
generalist who was doing doingthe campaign, uh, definition,
defining stuff, creating the,the creative, you know,
developing the creative,developing the content, um, now
that's all done by two, three,four different teams and who are
kind of, uh, working indifferent.
(16:13):
You know parts of the process.
How have you a I mean, am I theonly one who's seen this or you
just something you've seen?
And b, if you have seen it, howwould you think about trying to
minimize the impact of thatinefficiency that could creep in
in those organizations?
Speaker 2 (16:32):
Yeah, I mean getting everybody on the same step.
One is are we all aligned onthe strategy and the plan?
Is everybody intimately awareof what our KPIs and goals are?
And, as one team, you sit andlook at those every week and you
decide where do you need to,what do you need to stop doing
or what do you need to keep ondoing.
(16:52):
So alignment is critical, andit's not just alignment among
that full marketing team, butalso alignment with the sales
team as well.
And the other thing is and Iknow we'll probably get into it
in this discussion which is whatare some of the manual,
time-consuming, repetitive tasksthat are happening and taking
(17:16):
up all of your precious time?
That you could actually startusing technology and agentic
AI-type technology to speed upthe cycle.
And the part of I think thebiggest challenge with marketing
is you're.
There's so many.
Everybody works so closely andcollaboratively.
There's so manyinterdependencies and this has
(17:36):
to go on.
Then this this is a cycle or anatural workflow to what we do,
and I I'm pretty sure, becauseI've seen it firsthand and
worked with companies what can Iparallelize and do in concert?
Speaker 1 (17:51):
and ai can definitely help there yeah, okay, yeah, I
know that was just somethingthat popped in my head because
I've, like I I'm sure somewhere,like we've probably got a mix
of listeners and viewers whowork at smaller organizations
and probably, if they've neverexperienced what it's like in a
bigger team, how it can besometimes frustrating, like why
(18:15):
can't we get more done faster?
And it's kind of like we get inour own way in a lot of ways.
Speaker 2 (18:20):
Yeah, too many meetings to hypothesize what
went wrong without having accessto the right data set or
knowledge.
So the one thing I will add ona bigger team, that specialty
thing and everybody sort of getsdeep dive and narrower is that
a lot of times teams HubSpotisn't doing it for us and we
(18:43):
actually need to pull data fromall these other siloed systems,
whether it's our finance or oursales system, and so now we need
a data warehouse so they willget Snowflake, Stig, Tableau or
Looker on top of that, and thenwhat happens is you will slow
down because now you have tohire some data scientist or
engineer type person to gofigure out how to use that,
(19:06):
stand it up and share reports.
Speaker 1 (19:08):
So the more proficient you are with sharing
information that may be buriedin these silos and just being
active about that on aconsistent basis like it's the
same point about bringingeveryone together and along for
the ride and exposing them tothe critical information they
need to make decisions, to dotheir job yeah, I think in my
(19:29):
experience, if I were todiagnose most of the scenarios
where I've seen uh, inefficient,sort of across team say
campaign, just campaigndeployment, development
deployment a lot of it comesdown to A lot of it comes down
to reviewing approval processesthat are either unclear or
(19:57):
overly conservative, I guess isthe way I would put it, and
sometimes those come togetherright.
So by unclear I mean like who'sactually the one who can approve
it versus who can provide inputthat we may or may not
incorporate right Before youdeploy.
And I think that to me is like,if I could fix one thing in
(20:18):
most places is like try to givemore trust and then, uh, to the
teams to, to, to do the rightthing, and when something
happens, to get out the doorthat is maybe not ideal or has
something wrong or whatever, isthat you then provide grace to
those people?
It doesn't mean you give them atotal pass, right, it needs to
(20:40):
be a lesson from that.
But I think I think if therewere more teams who were
comfortable letting their teamsmove forward quickly, I think a
lot of that would go away.
Speaker 2 (20:53):
Yeah, a hundred percent.
I usually I'm alwaysencouraging team members to,
when they have an idea or asuggestion or a brainwave that
will help everybody likeproactively bring it up.
And I think you have toactively allow time and brain
space for that, because I thinkthe problem is you're always
(21:14):
running so hard and you'realways just executing.
You execute your report or youtry to figure out where you're
stuck or problems you rarelydevote time to.
Speaker 1 (21:24):
let's just have some white space to talk about
something and that's when likelet the almost let the quietest
person in the room speak up,because they probably have lots
of ideas that yeah you know soyeah, yeah, I don't have enough
time for that I, I agree, um,and in today's world I think
sometimes that's looked on like,oh, you're, like you're not
(21:45):
getting enough done right.
It's like this focus on beingbusy, which is not always the
best thing anyway, I mart, besmart, busy right.
I'm not prepared to talk aboutbrain science tonight, so today,
well, let's.
You mentioned AI and agenticstuff, so that's really, I think
(22:07):
, what we want to spend the restor much of the rest of the time
talking about.
But I brought it up when I wasintroducing you that you you
said to me when we spoke earlierthat AI will need HI, which you
said by human intelligence, andI know you wrote something
about this, I don't rememberwhere.
Speaker 2 (22:28):
It was a blog in the early part of the year Yep.
Speaker 1 (22:33):
So maybe we can put a link to it in the show notes
too, but why don't you walk usthrough what do you mean by that
?
Ai needs HI, and what is the HIpart of it?
Speaker 2 (22:42):
anyway, yeah, that's exactly right.
So the first of all, like Ithink everyone on the call today
is probably using ChatGPT,right?
Speaker 1 (23:02):
And I will admit that I have just like I was.
I've been slow to adopt it, butI've just literally.
I was just telling somebodyearlier today, just in the last.
Speaker 2 (23:08):
Two weeks, I'd say.
I've almost stopped usingGoogle for search.
Speaker 1 (23:10):
I've just started like take long right.
Well, it took me a long timecompared to others.
I'm sure they're like whatYou're just now getting to it,
but yeah, like it's I wastalking to to it.
Speaker 2 (23:16):
But yeah, um, like it's um.
I was talking to a formercolleague recently and it's sort
of like, um, yeah, it's mygo-to, like google search isn't
anymore because it gives you somuch more and it's if you're
stuck on something, it's a greatsort of memory memory jogger.
Or if you're frozen on writingsomething, like there's so many
(23:37):
different use cases and prompts.
So I'm assuming, like everybodyhas at least tried it on Now a
paid chat.
Gpt, then, is a little bit moresecure and a little bit more
advanced in its capabilities,and this is only one LLM that
exists in the market today.
So the good side of ChatGPT isthat it's actually telling us
(24:03):
that we can interface withsomething just using an English
prompt and get pretty rich,relevant response and answers.
That's awesome.
So I was like, oh my God, theinstantaneous of it is letting
us know that there's a promise.
If I really build out someagentic AI-driven workflows, I
(24:25):
can actually really get someinteresting work done.
The human intelligence part isthat it's just data and
technology.
It's not going to proactivelydo anything until you actually
design it.
So so you have to decide whatis it I want to get done, and
then how am I going to shape it.
(24:46):
What are the sort of thebeginning and the middle and the
end of this thing I want to do?
And that's in your controlright, your control right.
The other part of it is that youknow it will give you some
summary result or answer, butyou then have to decide what am
I going to do with thisadditional information or
content that I've just createdin a very fast fashion.
(25:09):
So where you want to place thatand how you want to distribute
or learn from it is entirelyyour decision.
So, first of all, I think AI isour, it's our helper, it's like
in our sidecar working with us.
It's hopefully going to be toldby us humans and learn over
(25:29):
time how to do all thoserepetitive manual things that I
don't want to spend time doing.
And then, over time it'll getmore sort of sophisticated and,
by its nature, get sort ofcomfortable with your use case.
Like Claude is very good atthat.
And it will learn your voice andyour tone when you start
(25:50):
producing content with that.
It's just an example of one way, but I'm like, I'm all for stop
doing these boring, repetitivetasks with our precious time and
use our time to do thestrategic, smarter work yeah, I
feel like, um, then I've had thesame conversation with other
people I feel like what?
Speaker 1 (26:12):
yes, this will.
It may completely replacecertain jobs, but I believe that
in most cases it won'tcompletely replace them.
It will change the nature ofthe jobs because, to your point,
(26:36):
say, a bunch of copy or textthat has a bunch of email
addresses in it, but they're allover, right, they're not in a
structured way.
I literally just put it into aLLM thing and said I'm going to
give you a bunch of text.
It has a bunch of emailaddresses.
Give me the email addressescomma separated, and it has
worked.
Bunch of email addresses.
(26:56):
Give me the email addressescomma separated, and it has
worked.
Beautiful, yeah, that wouldtake probably hours to do and
still, I probably still havemistakes yeah, are you like?
Speaker 2 (27:07):
oh, what was that thing in the pivot table and the
formula that I have to do inthe sheet and then yeah,
annoying and then you're doublechecking your work.
Speaker 1 (27:15):
Yeah, it's annoying, it's yeah, so, like I've, I'm
now like convinced, like this isif it can eliminate that kind
of crap.
And then, yeah, annoying.
And then you're double checkingyour work.
Yeah, it's annoying.
Yeah, so, like I'm now likeconvinced, like this is if it
can eliminate that kind of crapfrom my day to day.
And beautiful or where I'veused it now has been
(27:43):
no-transcript what we shouldlike, how we should organize it.
Was it great?
No, was it something that was astarting point?
Yeah, and it got me.
It got me started and that wasthe hardest part was just
getting started.
Speaker 2 (27:59):
Totally.
Yeah, I love it for that.
I'm currently using a copyediting writing tool, AI tool or
program.
It's called Copy Compass andit's part of a platform that's
provided by markov and it doesexactly that.
(28:20):
I can either upload somethingthat I might have written in
google doc or I can give it aprompt and then I can say like,
outline a long form article orblog, and then it'll create it
for me in a formatted way andthen I can start.
But I can highlight text and Ican say dive deeper, explain
(28:42):
more, make this assertive oractually cite and do references
and cross-linking.
It'll do spellcheck plagiarismstyle, does all of the above in
one simple platform, that'samazing.
One example right.
It's amazing one example.
Speaker 1 (28:59):
Right, that's amazing .
Speaker 2 (29:00):
It's like it just blows me away yeah it's, it's
awesome, like I actually enjoywriting myself.
So that is back to the pointabout like this is like getting
it faster, produced for you andformatted and laid out and
checked.
Then you come in as the humanand you put your personality
into that and you add in, but itmeans you can do it so much
(29:23):
faster.
Speaker 1 (29:23):
This reminds me you two.
You're totally off subject here, but I'll bring it back.
There's a movie that came outyears ago called Finding
Forrester.
Are you familiar with thatmovie?
Yeah, I am, yeah.
So this idea of starting withsomething else that somebody
else wrote and and but thenmaking it something of your own
is kind of what you're talkingabout.
Yes, because so now?
Speaker 2 (29:43):
that's like a little hint for all those people out
there listening if you haven'tseen finding forrester, go
listen, go watch it right, Ihave to go watch it again now,
but the there's a lot of talkand I I think I wrote a blog on
this recently a lot of talk andI think I wrote a blog on this
recently a lot of talk aboutlike.
people are like oh, traditionalSEO is dead and you know Google
has constantly changing theiralgorithms and their preferences
(30:08):
are for, like, clustering oftopics, not just one keyword.
So there's also things you haveto keep up with as an SEO type
marketer.
Speaker 1 (30:15):
So there's also things you have to keep up with
as an SEO type marketer.
Speaker 2 (30:17):
But the beauty of having a tool like that is it's
telling you is this AI produced?
Because you know you will getdinged if you're putting out
word salads on just trying tooptimize for one keyword.
So it's actually improving andmaking your quality better
(30:40):
because you can use theseagentic tools to tell you is
this, too, ai produced or isthis plagiarized?
um, that's really interestingthere's a couple of other little
repetitive tasks.
Use cases I'll give because Ialways like to give practical
advice, so like one examplewould be this so you can like
chat with your data which islike unstructured data, like
(31:02):
text, or you can chat with yourdata that's structured.
So let's say I output a CSVfile from my SEMrush or Ahrefs
where I'm doing analysis onkeyword research, seo, and you
can literally bring the CSV filein start text, chatting with it
(31:22):
and asking what keywords havevolume of like, a thousand, show
me all, produce it in a chartand then you will immediately
get insights and answers as towhere you should focus.
And that's just one littleexample of letting the AI do the
work.
And I didn't have to think oflike oh my God, I need to go
(31:43):
find somebody who does SQL to dosome analysis into some
warehouse.
You can literally just use alittle tool that's chatting with
your structured data.
Speaker 1 (31:53):
That's really interesting.
Yeah, I'm like I said, I'm, Iam.
I dipped my toes in very littlebit over the course of the last
12 months and I've reallystarted digging in.
I'm looking, I've been.
It's been interesting.
The other part that I've beenusing is some of the custom GPTs
that are I'm cheap, so I stillhave the free chat GPT, but
there's some great custom appsin there that are sort of
(32:16):
focused on certain things thathave been.
So far, most have been prettygood yes other than the image
generating ones.
I those I don't get a littlescary a little bit.
I think there's work to be donethere yeah, no, I agree with
you.
Speaker 2 (32:31):
Um, I think we're all very impressed with the speed
at which you can produce thosethings.
I think it is not veryauthentic looking, and I
actually worked at a companybefore that was all about visual
content licensing and thereality is that real photos and
(32:51):
videos exist on the internet.
We're all doing it every day onour phones and we're uploading
it to social.
So it's, it's in the in theworld online.
So if you can just access thatand then pull it into your
campaigns.
You're better off.
Speaker 1 (33:06):
So yeah, yes, yeah, I mean there's.
I did one where I was like Ineeded a, an image for a like a
blog post that I was doing forfor someone, and I said, yeah, I
gave it some ideas and hegenerated something that was
actually.
I was like, okay, that works.
But then I was looking atcloser and I was like it's
supposed to be like someone in akind of a.
(33:28):
It was a data, like a cdprelated thing, so there's data
and charts and stuff on thescreen.
It's kind of futuristic.
But somewhere randomly in themiddle was like a flower and I
was like why is that there?
the machine decided that yeah Imean, it was, it was the
strangest thing, um, butotherwise it was great, uh, but
(33:50):
I'm with you, like I've seensome really odd stuff come out
of those, those image generators, what.
So you've talked about a numberof scenarios here.
So, in general, where it goes,are there any ai tools that, um,
you know, marketing ops,listeners and viewers could be
(34:11):
thinking about?
This is how I could use aitools to help me in my
day-to-day job that are more opsfocused.
Speaker 2 (34:16):
Yeah.
So I will just say, likethere's the categorization of
like we just bubble it up for aminute, the categorization of
the things you can do with AI.
So to orient yourself, like amI actually doing some like
research or market research, andit could be more on a macro or
(34:38):
it could be on a very microlevel.
So a micro level might be I'mhelping the sales team really
understand that these targetaccounts we're going after and
do some enrichment andaugmentation of things like
industry size, public or not,geos, what's their claim or
(34:58):
their benefit or value.
So you're sort of like, as anops person, you're really
helping do a lift for that salesteam and so those are like a
macro level type research.
Or you might do something likescrape a competitor's website
and put it in a summary doc so Ican be really pointy about how
I'm going to differentiate Right.
(35:19):
So that's a research bucket.
Then we already sort of talkedabout some of the content
capabilities and then, from ademand perspective, like there
is workflow steps.
So you kind of touched on itearlier, which is like I have a
bunch of email addresses that Igot from a show or an event or
(35:40):
maybe some joint campaign I didwith somebody, and now I need to
like parse out what's thedomain of the company.
Now I need to parse out whatlike industry segment.
Now I need to parse out wherethey are, what size, what
revenue, all those enrichmentthings.
I need to parse out where theyare, what size, what revenue,
(36:00):
all those enrichment things.
So that is like a very valuableuse of like, just like you were
saying, oh my God, I'm doing itin Sheets or Google, and that
is a speeding up on theexecution of a highly targeted
campaign, for example.
So those kind of workflows, thetool that I'm using right now I
touched on earlier, which isit's a company called
(36:22):
MarkovMLcom and it, out of thebox like you literally
self-serve login, out of the boxyou can launch a specific thing
will do exactly that set oftasks.
All you have to do is sayhere's the file with the email
addresses, do the tasks and thenyou get the output and in less
(36:42):
than two minutes it's done right.
Speaker 1 (36:44):
Wow.
Speaker 2 (36:46):
And then I will.
I think most of your listenersthat are sort of leaning into
modern AI-type tools like Clayis the current darling know most
people are paying attention tothat.
Apollo is also a very good toolfor like, really getting the
right and clean emails fortarget and especially if you're
(37:09):
doing any outbound marketingtargeting certain accounts.
Um, the name of the game isthat.
Gone are the days where you dosort of a generic message to
this batch.
Right, you now have thecapabilities to be much more
targeted and personalized andnot like, oh, let's do like
(37:29):
three outbound emails a week orX number of things on your
calendar every week week or Xnumber of things on your
calendar every week.
Now you can be much moresophisticated and get higher
conversions and responses.
Speaker 1 (37:44):
That's really interesting.
Yeah, I can see that beinghelpful for our listeners as
well.
So you brought up Clay and youmentioned to me when we were
talking before about an exampleI don't know if you're able to
share it where it's saved timedoing outreach research.
(38:06):
I mean you touched on it there.
Can you share a little moreabout what that was like?
Speaker 2 (38:12):
In terms of the time it took to get it off the ground
.
Speaker 1 (38:17):
Well, I guess there's that too, but I'm just like
what?
What was it used to do thatotherwise would have been done
by through human capital right.
Speaker 2 (38:28):
Yeah, okay, that's good, good contrast.
So I mean we can take theexample of the play.
So I've worked with my sort offormer RevOps and MarketingOps
team members in Concert and thisis where, like, the human part
is coming in as well, right?
So we said, oh, we have theexisting customers that are
(38:52):
happy they're in this industryand segment.
How do we go find more of them?
So, like I think ZoomInfo andothers used to call it, where
are my lookalikes?
So you have to go figure outwhere are those ideal customers?
So that's kind of painful.
Like in the back in the day youwould like I'd mess around with
Zoom info.
Speaker 1 (39:13):
You had to go get D&D and yeah, all that stuff right.
Speaker 2 (39:16):
Yeah, a mix of tools, and then you put it into some
spreadsheet, and then you haveto augment the spreadsheet, and
then you have to check it, andthen you have to share it with
the sales team and then you haveto upload it and then you have
to go find all the emailaddresses on the right company
and hopefully the tools help you.
So in Clay's example, it is listbuilding for you, but doing it
(39:38):
in a very highly targeted way,and it is literally scraping
information like it's real time,so it's not like older data
that was maybe collected monthsor years ago.
It's literally now telling youthis company is doing this stuff
and here's the right personas.
Play is also going and findingthe persona information from
(40:03):
multiple sources, not just onesource, so they're sort of
casting a wider net.
So they'll go look at yourLinkedIn and probably find
information about you in otherplaces if they have a
relationship with Apollo, forexample, right.
And then it tells you all thatrich information.
And now you can actually besuper prescriptive and say
(40:27):
create a three-touch sequenceand you know top and tail my
messaging with these things buttailor it according to the human
and the company that they're at, based on what I want to engage
them with.
And that's where that humanpart comes in.
Speaker 1 (40:44):
So that's true, true, like truly personalized, as
opposed to just hello, michael,right, yes, I see you're at such
and such company yeah, exactly,and you know that it's very
sort of cut pasty, because whenyou read it it's like nothing to
do with me right.
Speaker 2 (40:59):
So, um, in the in the before times when you didn have
automation, you would ask yourBDRs hey, you got to go research
each of these target accountsand you have to then tailor your
message and you have to go intoyour outreach or your sales
loft or whatever you're using,and make those changes, and you
(41:21):
know how long that takes.
It's the right thing to do.
Speaker 1 (41:24):
It is the right thing to do.
Speaker 2 (41:25):
But they don't do it.
Speaker 1 (41:27):
No.
Speaker 2 (41:28):
Because it's a pain in the butt right.
Speaker 1 (41:30):
Sure, yeah, I mean I had a small inbound team at one
point in my career and we hadtemplates for large categories
of type of follow-up we'd havefor incoming leads and but we
always made sure that we tookthe time to tweak them to match
what we could tell about likedoing a little bit of research,
(41:52):
and I think that I've alwaysthought that that little bit of
extra effort reduced the volumethat we could do but was more
effective, and I would take thattrade off.
Speaker 2 (42:03):
Yes, quality over quantity, yes, so I think that
the, the delay tool, is allowingyou to strike that balance
right, you can get in, you canscale it easily and get the
right volume, but yourconversion rates are still way
higher and you know I have.
It takes some weeks to get thiswhole thing off the ground
(42:24):
because you are shaping it right, it's not.
You have to tell it what youwant it to do and then you have
to like you're not just going tosend out these.
It produced an email or messageand you want to put it in
LinkedIn.
You've got to validate it.
I'm like, I'm good with that,or put in your own little
personal touch.
So it definitely improves theconversions and I think that's
(42:46):
why people are adopting this,because it's working.
And you know, if we wanted it,we just didn't have the
technology that was enabling usto get there.
Speaker 1 (42:55):
Well, actually, I would argue that we did have the
technology.
We just didn't have the human,like we didn't have the time to
take advantage of it the way wewould want to.
Speaker 2 (43:03):
That's fair, but the technology allows us to speed it
up.
Speaker 1 (43:08):
Yes, totally agree with that and I like this.
I guess this is what we'retalking about when I hear about
agents, the AI agents andagentic stuff.
That's what we're talking about.
Right, it's like being able todo multiple steps with a little
bit of guidance on what you wantthe output to be.
Speaker 2 (43:23):
And I mean I'm sure you've been there where it's
like you realize that the teamis spending their precious time
on these repetitive things andyou're like, let's outsource
that.
And then you're like, okay, nowwe got to find a human that's
going to do this outsourced,boring work, and then we got to
pay them extra and then we gotto check their work.
Speaker 1 (43:49):
Yeah, yeah.
So it's yeah, yeah, for sure.
Um, well, let's maybe one morething, um, and then we can.
We can wrap up so one of thethings I'm bullish on and have
been bullish on about potentialfor ai to have an impact on is
actually somebody we talkedabout earlier reporting and
analytics, because I do believeit has been a very it's like a
very human capital intensivething.
Right, there's a lot of effortin addition, just expertise, and
(44:10):
so I am bullish on that.
Be able to help really drivesome great insights over time
quickly, whether it's faster, uh, be able to do more um
experiments or test morehypotheses, whatever it is right
.
I think any of those would bebeneficial.
Are you seeing any of that alsoout there in your experience
(44:34):
with?
I know you're doing a lot ofconsulting these days, so what's
your take on that area of AI'simpact?
Speaker 2 (44:41):
Yeah, the biggest challenge with and I actually
had a conversation with a headof an agency today who does all
ops and digital marketing andshe literally said I am
constantly wreck, I'm spendingall my day reconciling data, so
(45:02):
I call it like it's alwaysstranded and siloed, so it lives
inside Salesforce and it livesinside HubSpot, and as much as
you're syncing these systems,you still have to make sure it's
right.
So time and time again I would,at the end of every week or
even month or quarter, I wouldpull data out of both my you
(45:23):
know run the business systemsand then I put it into a sheet
and then I create pivots and addcolumns, vlookups and all that
yep yeah, and then I share itwith the ceo and the cfo and
whoever else needs to know, andit's a pain in the butt.
And then you, you're, you're inthere and you're actually making
charts out of it and so on.
So AI your question is like howis AI helping?
(45:47):
That AI is the I referenced itearlier in our conversation
where it's like I've got thatlovely CSV file.
Like you can turn most of thatstructured data into CSV.
I put it into my AI system.
I chat in English, chart me X,y, z or you know how many.
How many deals did we closelast quarter that were higher
(46:10):
than 150K, example?
Right, and it will instantlyproduce the answer put it in a
table and pick your chart styleand then output and share that.
And the thing, the beautifulthing about having english chat
with your structured data is youknow you would automatically do
(46:31):
that in your meeting anyway.
Now you can actually talk tothe system to do it.
And a question begets anotherquestion.
I've been I say this all thetime so, as soon as you, one
question and you get a nicelittle nugget and then you're
like, well, hang on a minute,how about this?
And that's the beauty of thereally pulling wisdom out of
(46:55):
your data, right?
If you can read the data andask the right questions, then
you can really start to beinformed and knowledgeable.
So I'm all about dive in playwith it.
Speaker 1 (47:07):
ask more questions, ask more questions that
iterative process is how youlike, really get the most out of
it right.
But it takes again, takes timeand effort and expertise, and so
, again, if we can make thatfaster but that this is a
another aside, but like this iswhy I whole knowledge that these
things one question leads toanother question leads to
(47:28):
another is why, when I go tosome place and I go, we need to
get dashboards.
No, you don't, you really don'tneed dashboards.
No, start with something small,exactly If that becomes
something you know you're goingto do regularly, because you
usually don't know, fine, makethat the beginnings of a
dashboard, then do the next one,but don't know, fine, make that
the beginnings of a dashboard,then do the next one, but don't
start with the idea of we'regoing to do this comprehensive
(47:49):
dashboard.
Speaker 2 (47:50):
I think you think I've never seen one of those
succeed exactly, and if youpresent a dashboard which has
like 12 things on it that giveyou a little insight into each
part, then you you have to like,if you're talking to your CFO,
for example, which the boardwill actually ask as well, then
they're just glazing overbecause it's too much, yeah,
(48:12):
whereas like the nuggets like Ilove your point about like, just
start with something small andI can't tell you how many
meetings I've been in and youpresent what you believe they
want to hear.
You go to the meeting with yournice, pretty and you know charts
and then, of course, they askmore questions and then you're
like well, I have to get back toyou on that because I didn't
(48:34):
produce that report and it slowseverything down yeah and or, or
, or maybe worse.
Speaker 1 (48:41):
You know you've got a dashboard with multiple reports
that are already tough toconsume but somebody goes that
one's not right, right, yeah,and, and then that that reduces
the credibility of all the otherones that may be.
Quote right.
I mean, first off, saying it'sright or wrong is I think it's
just a misunderstanding aboutwhat you're actually looking at.
(49:03):
But yeah, that's a whole that'sa podcast yeah, okay, well now.
So now I feel like I have got alot of catching up to do on all
these ai tools and getting someof the mundane stuff out of my
life.
Um, this has been great.
Thank you for for sharing.
If, if folks want to connectwith you or learn more about
what you're doing, or find yourpod, your, your blogs or
(49:26):
whatever you're doing, what'sthe best way for them to do that
?
Speaker 2 (49:30):
They can go on LinkedIn and hopefully Deirdre
Mann.
Speaker 1 (49:34):
It's a little tricky to spell but we'll make sure we
put a link to it somewhere alongthe way.
Speaker 2 (49:39):
Yeah, I do have a website.
It's called marketing enginetwocom, the number two, because
my my belief is that usuallyyounger companies they try it
and then they're like, okay,let's do this properly, b2.
And then you know just beingreally modern about your
approach.
So it's always a challenge.
(49:59):
So you can check me out thereor find me on LinkedIn and
message me directly.
Speaker 1 (50:04):
Sounds good.
I appreciate it, and you are,because we traded messages today
on the marketingopscom Slack,so you're there too, right?
Speaker 2 (50:11):
Absolutely.
Speaker 1 (50:12):
All right perfect.
Speaker 2 (50:12):
Thank you for all you do there, because I get lots of
learning from that.
Speaker 1 (50:17):
That's great.
Well, we appreciate it.
Thank you again.
Thanks to all of our listenersout there for continuing to
support us and sharing yourideas and topics on guests and,
as always, if you want to be aguest or have a topic you want
to share with us to cover, feelfree to reach out to Naomi, Mike
or me, and we'd be glad tofollow up with that Till next
(50:37):
time.
Bye, everybody.
Hello everyone, welcome to another episode of
OpsCast brought to you byMarketingOpscom, powered by the
MoPros out there.
I'm your host, michael Hartman,joined today by well, just our
guest who we'll get to Joiningme today is Deirdre Mahon.
Deirdre is currently a productmarketing and growth advisor
with Superset, advising theirportfolio companies.
She is a full-stack marketingleader with hands-on experience
(00:22):
across growth startups andpublic companies.
Particularly strong atexecuting against strategy and
driving the tactics to exceedgoals.
She has been instrumental inthe creation of new market
category segments, includingobservability, real-time data,
movement, cloud consumption,analytics and big data.
And, as she said this is not memaking this up, she said this
(00:42):
she's probably been acquiredfive times.
So, deidre, thank you forjoining us today.
Speaker 2 (00:47):
Thank you for having me and lovely to be with you,
Michael.
Speaker 1 (00:50):
Yeah, yeah.
Well, we were enjoying a nicechat about rugby before this, so
we better get back to business.
So, for our listeners, we'regoing to talk a little bit about
a number of different things,but really it's going to be
centered around the idea that AIneeds human intelligence.
We'll get to there in a bit.
(01:11):
So, deirdre, during your career, you have worked mostly in
marketing and leadership rolesat early stage companies, some
larger companies, and you'veworked closely with ops
professionals to stand up thatpart of the marketing function.
So, you know, I think it wouldbe useful for our listeners,
(01:31):
before we get into the maintopic, to kind of get your
perspective as a market leaderon what you see.
As you know, what is abenchmark kind of marketing
operations function compared toothers?
What do you look for in thoseroles?
Speaker 2 (01:47):
Yes, by that question I'm assuming, mike, you're
talking about what is a goodstarting out marketing
operations look like from ahuman side and otherwise right.
Speaker 1 (02:01):
I mean, I think you could take it wherever you want.
But yeah, I'm thinking about, Ithink there's going to be
probably two main categories,right.
One is starting it from scratch, which I think some people get
to do most, don't, uh?
And if they do do it, thenthey're, you know, they're
juggling everything, yep um, andthen maybe a little bit of
(02:22):
overtime.
And then others who areprobably in sort of a role
similar to mine.
Very often, even if I wasstarting something, it was
partially in progress, or Iinherited something-series A,
and you're right, sometimes youknow somebody else and they
could be a sales or somego-to-market human has, like, at
least got HubSpot licensed, butI mean that's just the very
(03:00):
sort of minimal viable.
Speaker 2 (03:02):
So when I think of a true functioning marketing
operations practice, I thinkfirst things is the engine.
So what is your engine thatyou're actually going to execute
?
And the obvious thing is what'syour marketing automation
system?
Most people on the call todayare probably either using
(03:25):
HubSpot or a close competitor,hubspot.
We love them because they'veactually helped younger
companies get on the engineplatform as quickly as possible
and not super expensive.
So I personally do like HubSpot.
Then the second thing is OK, dowe have an agreed upon set of
(03:50):
process steps?
And you know, if you have asales team in place or like some
small team that's actuallyactively selling and engaging
customers, then that's awesome.
So you have a partner thatyou're going to work on defining
those process steps.
And one of the most importantthings to get right out of the
(04:12):
gate is what's your model?
So are you a sales led?
Are you product led?
Are you hybrid?
Right?
So your process steps should bedirectly mapping to how people
engage and buy.
And then the last third pieceand then I'll touch on the
people side of it is.
The last piece is do you knowwhat you're actually going to
(04:33):
measure and do you have anagreed upon set of KPIs or goals
?
And how are you going to divideand conquer and co-own and have
accountability for thosenumbers?
So that's where, hopefully,you're going to sit down with
your sales leader and yourperson in finance.
Based on the model, figure outwhat those numbers look like
(04:55):
okay, I was gonna.
Speaker 1 (04:57):
I was gonna ask you if you met goals and metrics for
the marketing I've seen, formarketing or for the
organization, maybe the revenueengine, if you will.
But I think you just by whatyou just said there, answered it
that it's really more foreither all marketing or all of
the marketing sales functions.
Speaker 2 (05:17):
Yeah, Just to touch on your point about you know
goals and metrics for the opsteam.
A lot of times, especially inthe early days when you're
standing up the engine a lot oftimes it's like, okay, can you
deploy and get this new tool orsystem or process in place by x?
date train and enable otherpeople on the team to be
(05:40):
comfortable and proficient withusing it.
And also you.
You know you immediately assoon as you're buying any tech
system, you're immediately onbudget topic.
So as a team, have we carvedout enough budget to do all that
system engine and trackingstuff?
And sometimes that's like, oh,it's third level of importance,
(06:06):
because oftentimes people arethinking like, how are we going
to put money into campaigns andprograms so we can get the word
out and start campaigning andyou know outbending?
And you forget oh, don't, weneed some budget for the tools
and the process steps and theonboarding.
Need some budget for the toolsand the process steps.
(06:27):
On the onboarding.
And I will say from a peopleside, when you're young and you
have a very small team andthere's maybe two people, doing
this who have other jobs as well, by the way, that you often are
like a little drowning eventhough the rest a lot of times,
sadly, the rest of theorganization is going.
You're just building processand engine.
When are you going to actuallydo some marketing Right?
(06:47):
So you really have to sort ofstrike that balance and I highly
recommend all the listeners tolike invest early in the engine
because you don't want to begoing back and fixing it later.
Get the foundation right andthen, from a people standpoint,
pull in some specialtycontractors that can, you know,
(07:07):
be a catalyst and fast forward.
Just make sure that you pickthe right people.
You know you don't want tobreak your budget, but it does
tremendously help if you have ayounger, newer team that doesn't
know all the moving parts toall the tech, that doesn't know
all the moving parts to all thetech.
Speaker 1 (07:24):
So just curious, going back to the people,
especially as starting up a newfunction, would you look for
someone who's a little more of ageneralist and maybe not an
expert in, say, a marketingautomation platform per se, but
has the skills to do it?
Or do you look for I need theperson who's an expert in just
(07:47):
use HubSpot as an example, sinceyou brought it up but the
expert in HubSpot who haspotential to be able to do the
other things and support themarketing teams and be able to
be an advisor to those teams.
Speaker 2 (08:01):
Yes, good question.
My bent is towards thegeneralist because when you're a
small team wearing lots of hats, if you hire only a specialty
person that knows only one thing, then you'll end up you know
you yourself will end up pickingup all the other pieces.
So I want to hire somebody who'sa generalist.
(08:21):
If that person has someknowledge and experience with
HubSpot out the gate and has anappetite to learn and grow,
which, if I'm honest, mostmarketing operations folks that
I've worked with in the past.
They have ambitious careers, sothat's awesome.
I have worked with folks thatare really comfortable with data
(08:43):
and really comfortable with inthe weeds.
That that's more important.
If that individual iscomfortable doing that and then
you're on a good path and easilyshow them the growth trajectory
out of that.
They're not going to be in theweeds forever, but they have to
be comfortable with that.
Yeah.
Speaker 1 (09:02):
I've rarely had the opportunity to really build a
team from the ground up and Ithink early on I would have said
I need somebody who can run theengine, call the marketing
automation platform, and be ableto fulfill needs like building
emails, getting them deployedand that kind of stuff.
And then after that I wouldprobably want someone who can
(09:25):
take some of that like split itup a little bit, but I think
I've gotten to the point whereI'm glad you brought up the
analytics piece, because I thinkI've shifted a little bit that
I think a second hire would needto.
I'd want somebody who reallycould understand data analytics
and help us get insights,because I think there's a lot of
potential leverage from thatthat you don't get if all you're
(09:46):
doing is executing on campaignsand not to say those aren't
important.
But um, at least up untilrecently and we've had some
interesting conversations withguests and vendors and stuff
like that but um, I'm just likeI just that reporting analytics
piece is such a unique.
I don't think there's a deepenough skillset in the pool of
(10:07):
talent out there, nor do I thinkit's a.
It can be just built on top ofreporting tools, right?
I think there's an effortthat's required to do it well.
So it's interesting that youbring that up to that as well.
So that's just because I dothink it's a really, really
valuable one that's underappreciated by other marketing
(10:28):
leaders I've run into yeah, 100the.
Speaker 2 (10:32):
The reality is that marketing has a ton of data,
right?
Our systems are naturallygathering it all, and that's
from like website traffic toconversions and engagement on
your HubSpot embedded forms, tothen how it downstream converts
in the funnel.
And first of all you have toknow the process and the model,
(10:55):
and then you have to startcollecting the data and start
having some benchmark orbaseline, and only then can you
figure out where gaps are orstuckness or like, without
having exposure to the data anda comfort around how to even
read and deduce from the data,then you really can't move
(11:17):
forward.
so, and that's number one andnumber two, like the marketing
team has to get reallycomfortable with that and then
you have to be able to sort oftranslate that to other parts of
the business so they understand, because if you drown them with
too much information, too muchdata, that just glays over.
Yeah.
So you have to pull back andsay, okay, here's the most
(11:37):
important things sales leader orCEO.
So getting really comfortablewith data and understanding it
is critical in my opinion.
Speaker 1 (11:47):
Yeah, well, and I like the second part of
translating or the storytellingpart of it, exactly Doing the
analytics, all right.
So you asked me the questionback at me about what kind of
organization we're talking about.
You talked a lot now about thestartup world, like do you, in
your experience with largerorganizations, like, what do you
(12:09):
see that's different in thoseearly stage kind of companies
and what they're doing from amarketing, marketing app
standpoint, and those ones thatare, say, medium, larger, maybe
even enterprise?
Speaker 2 (12:20):
Yeah.
So the smaller stage is likeyou're going to get your MVP
right, get the engine built, andthen you start to execute and
you start to measure.
And I always think of it in aone-two punch way you invest,
you execute, you measure, andthe shrinking of the time
between those two things iscritical so you can move forward
(12:42):
.
Two things is critical so youcan move forward.
When you're further along,hopefully, you know what a good
sales cycle looks like and yourgoal is to get it repeatable and
to make sure everybody's on thesame page with your
understanding of here's what arepeatable looks like.
And then you're probably goingto be like okay, how do we scale
(13:03):
up?
So, from a marketingperspective, your operations
teams are, besides all theexecution that they have to do
and focusing in on the thingsthat work and measure.
You're now figuring out how do Ido it at scale, because it's
always up into the right growthyes um, and so in case, what
(13:25):
usually happens is you actuallystop being more of a generalist
and you start being more of aspecialty expert.
So there's, like I've seendifferent ways to skin this, but
sometimes you have a dedicatedteam that is only focused on top
of funnel, and then you haveanother team that's focused on
like sort of of funnel, and thenyou have another team that's
(13:45):
focused on like sort of midfunnel.
And then maybe even somebodywho's working really closely
with product.
If you're more of a self-servePOG and you're more sort of in
the bottom of the funnel, whichis all about usage and adoption
and expansion and renewals,right, Very classic POG.
Speaker 1 (14:04):
Some customer marketing in some places would
be yeah, exactly.
Speaker 2 (14:08):
Because you're further along and you have
existing base and you have tokeep them happy and learn from
that and upsell, cross-sell andso on, references and et cetera.
So you end up being like I likethis sort of demarcation between
am I a generalist or aspecialist.
So now you're probably more ofa specialty and you're focused
on a particular part of thebusiness or maybe a particular
(14:30):
segment, or you're working withthe enterprise sales team
directly to help them, versusmaybe self-serve side of the
house.
So ultimately, at the end ofthe day, you're honing and
tuning what's working so thatany additional marketing dollars
that you're pouring in, thatyou're being really efficient
(14:52):
and spending where you know thatthere's going to be a growth at
those different buying cyclestages.
Now you know every week whenyou do your weekly business
review.
Now you know what is yourexpected goals and KPIs and how
(15:13):
are you mapping each week tothose.
So each of those dedicatedteams can really get much closer
to the problem and makerecommendations on what to do to
engage, whether it's top,middle or bottom of funnel.
Speaker 1 (15:29):
So one of the things I've experienced and I don't
know if you've seen this is Ithink this is a real challenge.
As teams get larger, you think,I think intuitively, you're
like oh, we can multiply theamount we can get done, but I've
found that that multiplier ismuch smaller than we think it is
because there's be therebecomes inefficiencies in how
(15:51):
things get done, because nowyou're, instead of you had the
generalist who was doing doingthe campaign, uh, definition,
defining stuff, creating the,the creative, you know,
developing the creative,developing the content, um, now
that's all done by two, three,four different teams and who are
kind of, uh, working indifferent.
(16:13):
You know parts of the process.
How have you a I mean, am I theonly one who's seen this or you
just something you've seen?
And b, if you have seen it, howwould you think about trying to
minimize the impact of thatinefficiency that could creep in
in those organizations?
Speaker 2 (16:32):
Yeah, I mean getting everybody on the same step.
One is are we all aligned onthe strategy and the plan?
Is everybody intimately awareof what our KPIs and goals are?
And, as one team, you sit andlook at those every week and you
decide where do you need to,what do you need to stop doing
or what do you need to keep ondoing.
(16:52):
So alignment is critical, andit's not just alignment among
that full marketing team, butalso alignment with the sales
team as well.
And the other thing is and Iknow we'll probably get into it
in this discussion which is whatare some of the manual,
time-consuming, repetitive tasksthat are happening and taking
(17:16):
up all of your precious time?
That you could actually startusing technology and agentic
AI-type technology to speed upthe cycle.
And the part of I think thebiggest challenge with marketing
is you're.
There's so many.
Everybody works so closely andcollaboratively.
There's so manyinterdependencies and this has
(17:36):
to go on.
Then this this is a cycle or anatural workflow to what we do,
and I I'm pretty sure, becauseI've seen it firsthand and
worked with companies what can Iparallelize and do in concert?
Speaker 1 (17:51):
and ai can definitely help there yeah, okay, yeah, I
know that was just somethingthat popped in my head because
I've, like I I'm sure somewhere,like we've probably got a mix
of listeners and viewers whowork at smaller organizations
and probably, if they've neverexperienced what it's like in a
bigger team, how it can besometimes frustrating, like why
(18:15):
can't we get more done faster?
And it's kind of like we get inour own way in a lot of ways.
Speaker 2 (18:20):
Yeah, too many meetings to hypothesize what
went wrong without having accessto the right data set or
knowledge.
So the one thing I will add ona bigger team, that specialty
thing and everybody sort of getsdeep dive and narrower is that
a lot of times teams HubSpotisn't doing it for us and we
(18:43):
actually need to pull data fromall these other siloed systems,
whether it's our finance or oursales system, and so now we need
a data warehouse so they willget Snowflake, Stig, Tableau or
Looker on top of that, and thenwhat happens is you will slow
down because now you have tohire some data scientist or
engineer type person to gofigure out how to use that,
(19:06):
stand it up and share reports.
Speaker 1 (19:08):
So the more proficient you are with sharing
information that may be buriedin these silos and just being
active about that on aconsistent basis like it's the
same point about bringingeveryone together and along for
the ride and exposing them tothe critical information they
need to make decisions, to dotheir job yeah, I think in my
(19:29):
experience, if I were todiagnose most of the scenarios
where I've seen uh, inefficient,sort of across team say
campaign, just campaigndeployment, development
deployment a lot of it comesdown to A lot of it comes down
to reviewing approval processesthat are either unclear or
(19:57):
overly conservative, I guess isthe way I would put it, and
sometimes those come togetherright.
So by unclear I mean like who'sactually the one who can approve
it versus who can provide inputthat we may or may not
incorporate right Before youdeploy.
And I think that to me is like,if I could fix one thing in
(20:18):
most places is like try to givemore trust and then, uh, to the
teams to, to, to do the rightthing, and when something
happens, to get out the doorthat is maybe not ideal or has
something wrong or whatever, isthat you then provide grace to
those people?
It doesn't mean you give them atotal pass, right, it needs to
(20:40):
be a lesson from that.
But I think I think if therewere more teams who were
comfortable letting their teamsmove forward quickly, I think a
lot of that would go away.
Speaker 2 (20:53):
Yeah, a hundred percent.
I usually I'm alwaysencouraging team members to,
when they have an idea or asuggestion or a brainwave that
will help everybody likeproactively bring it up.
And I think you have toactively allow time and brain
space for that, because I thinkthe problem is you're always
(21:14):
running so hard and you'realways just executing.
You execute your report or youtry to figure out where you're
stuck or problems you rarelydevote time to.
Speaker 1 (21:24):
let's just have some white space to talk about
something and that's when likelet the almost let the quietest
person in the room speak up,because they probably have lots
of ideas that yeah you know soyeah, yeah, I don't have enough
time for that I, I agree, um,and in today's world I think
sometimes that's looked on like,oh, you're, like you're not
(21:45):
getting enough done right.
It's like this focus on beingbusy, which is not always the
best thing anyway, I mart, besmart, busy right.
I'm not prepared to talk aboutbrain science tonight, so today,
well, let's.
You mentioned AI and agenticstuff, so that's really, I think
(22:07):
, what we want to spend the restor much of the rest of the time
talking about.
But I brought it up when I wasintroducing you that you you
said to me when we spoke earlierthat AI will need HI, which you
said by human intelligence, andI know you wrote something
about this, I don't rememberwhere.
Speaker 2 (22:28):
It was a blog in the early part of the year Yep.
Speaker 1 (22:33):
So maybe we can put a link to it in the show notes
too, but why don't you walk usthrough what do you mean by that
?
Ai needs HI, and what is the HIpart of it?
Speaker 2 (22:42):
anyway, yeah, that's exactly right.
So the first of all, like Ithink everyone on the call today
is probably using ChatGPT,right?
Speaker 1 (23:02):
And I will admit that I have just like I was.
I've been slow to adopt it, butI've just literally.
I was just telling somebodyearlier today, just in the last.
Speaker 2 (23:08):
Two weeks, I'd say.
I've almost stopped usingGoogle for search.
Speaker 1 (23:10):
I've just started like take long right.
Well, it took me a long timecompared to others.
I'm sure they're like whatYou're just now getting to it,
but yeah, like it's I wastalking to to it.
Speaker 2 (23:16):
But yeah, um, like it's um.
I was talking to a formercolleague recently and it's sort
of like, um, yeah, it's mygo-to, like google search isn't
anymore because it gives you somuch more and it's if you're
stuck on something, it's a greatsort of memory memory jogger.
Or if you're frozen on writingsomething, like there's so many
(23:37):
different use cases and prompts.
So I'm assuming, like everybodyhas at least tried it on Now a
paid chat.
Gpt, then, is a little bit moresecure and a little bit more
advanced in its capabilities,and this is only one LLM that
exists in the market today.
So the good side of ChatGPT isthat it's actually telling us
(24:03):
that we can interface withsomething just using an English
prompt and get pretty rich,relevant response and answers.
That's awesome.
So I was like, oh my God, theinstantaneous of it is letting
us know that there's a promise.
If I really build out someagentic AI-driven workflows, I
(24:25):
can actually really get someinteresting work done.
The human intelligence part isthat it's just data and
technology.
It's not going to proactivelydo anything until you actually
design it.
So so you have to decide whatis it I want to get done, and
then how am I going to shape it.
(24:46):
What are the sort of thebeginning and the middle and the
end of this thing I want to do?
And that's in your controlright, your control right.
The other part of it is that youknow it will give you some
summary result or answer, butyou then have to decide what am
I going to do with thisadditional information or
content that I've just createdin a very fast fashion.
(25:09):
So where you want to place thatand how you want to distribute
or learn from it is entirelyyour decision.
So, first of all, I think AI isour, it's our helper, it's like
in our sidecar working with us.
It's hopefully going to be toldby us humans and learn over
(25:29):
time how to do all thoserepetitive manual things that I
don't want to spend time doing.
And then, over time it'll getmore sort of sophisticated and,
by its nature, get sort ofcomfortable with your use case.
Like Claude is very good atthat.
And it will learn your voice andyour tone when you start
(25:50):
producing content with that.
It's just an example of one way, but I'm like, I'm all for stop
doing these boring, repetitivetasks with our precious time and
use our time to do thestrategic, smarter work yeah, I
feel like, um, then I've had thesame conversation with other
people I feel like what?
Speaker 1 (26:12):
yes, this will.
It may completely replacecertain jobs, but I believe that
in most cases it won'tcompletely replace them.
It will change the nature ofthe jobs because, to your point,
(26:36):
say, a bunch of copy or textthat has a bunch of email
addresses in it, but they're allover, right, they're not in a
structured way.
I literally just put it into aLLM thing and said I'm going to
give you a bunch of text.
It has a bunch of emailaddresses.
Give me the email addressescomma separated, and it has
worked.
Bunch of email addresses.
(26:56):
Give me the email addressescomma separated, and it has
worked.
Beautiful, yeah, that wouldtake probably hours to do and
still, I probably still havemistakes yeah, are you like?
Speaker 2 (27:07):
oh, what was that thing in the pivot table and the
formula that I have to do inthe sheet and then yeah,
annoying and then you're doublechecking your work.
Speaker 1 (27:15):
Yeah, it's annoying, it's yeah, so, like I've, I'm
now like convinced, like this isif it can eliminate that kind
of crap.
And then, yeah, annoying.
And then you're double checkingyour work.
Yeah, it's annoying.
Yeah, so, like I'm now likeconvinced, like this is if it
can eliminate that kind of crapfrom my day to day.
And beautiful or where I'veused it now has been
(27:43):
no-transcript what we shouldlike, how we should organize it.
Was it great?
No, was it something that was astarting point?
Yeah, and it got me.
It got me started and that wasthe hardest part was just
getting started.
Speaker 2 (27:59):
Totally.
Yeah, I love it for that.
I'm currently using a copyediting writing tool, AI tool or
program.
It's called Copy Compass andit's part of a platform that's
provided by markov and it doesexactly that.
(28:20):
I can either upload somethingthat I might have written in
google doc or I can give it aprompt and then I can say like,
outline a long form article orblog, and then it'll create it
for me in a formatted way andthen I can start.
But I can highlight text and Ican say dive deeper, explain
(28:42):
more, make this assertive oractually cite and do references
and cross-linking.
It'll do spellcheck plagiarismstyle, does all of the above in
one simple platform, that'samazing.
One example right.
It's amazing one example.
Speaker 1 (28:59):
Right, that's amazing .
Speaker 2 (29:00):
It's like it just blows me away yeah it's, it's
awesome, like I actually enjoywriting myself.
So that is back to the pointabout like this is like getting
it faster, produced for you andformatted and laid out and
checked.
Then you come in as the humanand you put your personality
into that and you add in, but itmeans you can do it so much
(29:23):
faster.
Speaker 1 (29:23):
This reminds me you two.
You're totally off subject here, but I'll bring it back.
There's a movie that came outyears ago called Finding
Forrester.
Are you familiar with thatmovie?
Yeah, I am, yeah.
So this idea of starting withsomething else that somebody
else wrote and and but thenmaking it something of your own
is kind of what you're talkingabout.
Yes, because so now?
Speaker 2 (29:43):
that's like a little hint for all those people out
there listening if you haven'tseen finding forrester, go
listen, go watch it right, Ihave to go watch it again now,
but the there's a lot of talkand I I think I wrote a blog on
this recently a lot of talk andI think I wrote a blog on this
recently a lot of talk aboutlike.
people are like oh, traditionalSEO is dead and you know Google
has constantly changing theiralgorithms and their preferences
(30:08):
are for, like, clustering oftopics, not just one keyword.
So there's also things you haveto keep up with as an SEO type
marketer.
Speaker 1 (30:15):
So there's also things you have to keep up with
as an SEO type marketer.
Speaker 2 (30:17):
But the beauty of having a tool like that is it's
telling you is this AI produced?
Because you know you will getdinged if you're putting out
word salads on just trying tooptimize for one keyword.
So it's actually improving andmaking your quality better
(30:40):
because you can use theseagentic tools to tell you is
this, too, ai produced or isthis plagiarized?
um, that's really interestingthere's a couple of other little
repetitive tasks.
Use cases I'll give because Ialways like to give practical
advice, so like one examplewould be this so you can like
chat with your data which islike unstructured data, like
(31:02):
text, or you can chat with yourdata that's structured.
So let's say I output a CSVfile from my SEMrush or Ahrefs
where I'm doing analysis onkeyword research, seo, and you
can literally bring the CSV filein start text, chatting with it
(31:22):
and asking what keywords havevolume of like, a thousand, show
me all, produce it in a chartand then you will immediately
get insights and answers as towhere you should focus.
And that's just one littleexample of letting the AI do the
work.
And I didn't have to think oflike oh my God, I need to go
(31:43):
find somebody who does SQL to dosome analysis into some
warehouse.
You can literally just use alittle tool that's chatting with
your structured data.
Speaker 1 (31:53):
That's really interesting.
Yeah, I'm like I said, I'm, Iam.
I dipped my toes in very littlebit over the course of the last
12 months and I've reallystarted digging in.
I'm looking, I've been.
It's been interesting.
The other part that I've beenusing is some of the custom GPTs
that are I'm cheap, so I stillhave the free chat GPT, but
there's some great custom appsin there that are sort of
(32:16):
focused on certain things thathave been.
So far, most have been prettygood yes other than the image
generating ones.
I those I don't get a littlescary a little bit.
I think there's work to be donethere yeah, no, I agree with
you.
Speaker 2 (32:31):
Um, I think we're all very impressed with the speed
at which you can produce thosethings.
I think it is not veryauthentic looking, and I
actually worked at a companybefore that was all about visual
content licensing and thereality is that real photos and
(32:51):
videos exist on the internet.
We're all doing it every day onour phones and we're uploading
it to social.
So it's, it's in the in theworld online.
So if you can just access thatand then pull it into your
campaigns.
You're better off.
Speaker 1 (33:06):
So yeah, yes, yeah, I mean there's.
I did one where I was like Ineeded a, an image for a like a
blog post that I was doing forfor someone, and I said, yeah, I
gave it some ideas and hegenerated something that was
actually.
I was like, okay, that works.
But then I was looking atcloser and I was like it's
supposed to be like someone in akind of a.
(33:28):
It was a data, like a cdprelated thing, so there's data
and charts and stuff on thescreen.
It's kind of futuristic.
But somewhere randomly in themiddle was like a flower and I
was like why is that there?
the machine decided that yeah Imean, it was, it was the
strangest thing, um, butotherwise it was great, uh, but
(33:50):
I'm with you, like I've seensome really odd stuff come out
of those, those image generators, what.
So you've talked about a numberof scenarios here.
So, in general, where it goes,are there any ai tools that, um,
you know, marketing ops,listeners and viewers could be
(34:11):
thinking about?
This is how I could use aitools to help me in my
day-to-day job that are more opsfocused.
Speaker 2 (34:16):
Yeah.
So I will just say, likethere's the categorization of
like we just bubble it up for aminute, the categorization of
the things you can do with AI.
So to orient yourself, like amI actually doing some like
research or market research, andit could be more on a macro or
(34:38):
it could be on a very microlevel.
So a micro level might be I'mhelping the sales team really
understand that these targetaccounts we're going after and
do some enrichment andaugmentation of things like
industry size, public or not,geos, what's their claim or
(34:58):
their benefit or value.
So you're sort of like, as anops person, you're really
helping do a lift for that salesteam and so those are like a
macro level type research.
Or you might do something likescrape a competitor's website
and put it in a summary doc so Ican be really pointy about how
I'm going to differentiate Right.
(35:19):
So that's a research bucket.
Then we already sort of talkedabout some of the content
capabilities and then, from ademand perspective, like there
is workflow steps.
So you kind of touched on itearlier, which is like I have a
bunch of email addresses that Igot from a show or an event or
(35:40):
maybe some joint campaign I didwith somebody, and now I need to
like parse out what's thedomain of the company.
Now I need to parse out whatlike industry segment.
Now I need to parse out wherethey are, what size, what
revenue, all those enrichmentthings.
I need to parse out where theyare, what size, what revenue,
(36:00):
all those enrichment things.
So that is like a very valuableuse of like, just like you were
saying, oh my God, I'm doing itin Sheets or Google, and that
is a speeding up on theexecution of a highly targeted
campaign, for example.
So those kind of workflows, thetool that I'm using right now I
touched on earlier, which isit's a company called
(36:22):
MarkovMLcom and it, out of thebox like you literally
self-serve login, out of the boxyou can launch a specific thing
will do exactly that set oftasks.
All you have to do is sayhere's the file with the email
addresses, do the tasks and thenyou get the output and in less
(36:42):
than two minutes it's done right.
Speaker 1 (36:44):
Wow.
Speaker 2 (36:46):
And then I will.
I think most of your listenersthat are sort of leaning into
modern AI-type tools like Clayis the current darling know most
people are paying attention tothat.
Apollo is also a very good toolfor like, really getting the
right and clean emails fortarget and especially if you're
(37:09):
doing any outbound marketingtargeting certain accounts.
Um, the name of the game isthat.
Gone are the days where you dosort of a generic message to
this batch.
Right, you now have thecapabilities to be much more
targeted and personalized andnot like, oh, let's do like
(37:29):
three outbound emails a week orX number of things on your
calendar every week week or Xnumber of things on your
calendar every week.
Now you can be much moresophisticated and get higher
conversions and responses.
Speaker 1 (37:44):
That's really interesting.
Yeah, I can see that beinghelpful for our listeners as
well.
So you brought up Clay and youmentioned to me when we were
talking before about an exampleI don't know if you're able to
share it where it's saved timedoing outreach research.
(38:06):
I mean you touched on it there.
Can you share a little moreabout what that was like?
Speaker 2 (38:12):
In terms of the time it took to get it off the ground
.
Speaker 1 (38:17):
Well, I guess there's that too, but I'm just like
what?
What was it used to do thatotherwise would have been done
by through human capital right.
Speaker 2 (38:28):
Yeah, okay, that's good, good contrast.
So I mean we can take theexample of the play.
So I've worked with my sort offormer RevOps and MarketingOps
team members in Concert and thisis where, like, the human part
is coming in as well, right?
So we said, oh, we have theexisting customers that are
(38:52):
happy they're in this industryand segment.
How do we go find more of them?
So, like I think ZoomInfo andothers used to call it, where
are my lookalikes?
So you have to go figure outwhere are those ideal customers?
So that's kind of painful.
Like in the back in the day youwould like I'd mess around with
Zoom info.
Speaker 1 (39:13):
You had to go get D&D and yeah, all that stuff right.
Speaker 2 (39:16):
Yeah, a mix of tools, and then you put it into some
spreadsheet, and then you haveto augment the spreadsheet, and
then you have to check it, andthen you have to share it with
the sales team and then you haveto upload it and then you have
to go find all the emailaddresses on the right company
and hopefully the tools help you.
So in Clay's example, it is listbuilding for you, but doing it
(39:38):
in a very highly targeted way,and it is literally scraping
information like it's real time,so it's not like older data
that was maybe collected monthsor years ago.
It's literally now telling youthis company is doing this stuff
and here's the right personas.
Play is also going and findingthe persona information from
(40:03):
multiple sources, not just onesource, so they're sort of
casting a wider net.
So they'll go look at yourLinkedIn and probably find
information about you in otherplaces if they have a
relationship with Apollo, forexample, right.
And then it tells you all thatrich information.
And now you can actually besuper prescriptive and say
(40:27):
create a three-touch sequenceand you know top and tail my
messaging with these things buttailor it according to the human
and the company that they're at, based on what I want to engage
them with.
And that's where that humanpart comes in.
Speaker 1 (40:44):
So that's true, true, like truly personalized, as
opposed to just hello, michael,right, yes, I see you're at such
and such company yeah, exactly,and you know that it's very
sort of cut pasty, because whenyou read it it's like nothing to
do with me right.
Speaker 2 (40:59):
So, um, in the in the before times when you didn have
automation, you would ask yourBDRs hey, you got to go research
each of these target accountsand you have to then tailor your
message and you have to go intoyour outreach or your sales
loft or whatever you're using,and make those changes, and you
(41:21):
know how long that takes.
It's the right thing to do.
Speaker 1 (41:24):
It is the right thing to do.
Speaker 2 (41:25):
But they don't do it.
Speaker 1 (41:27):
No.
Speaker 2 (41:28):
Because it's a pain in the butt right.
Speaker 1 (41:30):
Sure, yeah, I mean I had a small inbound team at one
point in my career and we hadtemplates for large categories
of type of follow-up we'd havefor incoming leads and but we
always made sure that we tookthe time to tweak them to match
what we could tell about likedoing a little bit of research,
(41:52):
and I think that I've alwaysthought that that little bit of
extra effort reduced the volumethat we could do but was more
effective, and I would take thattrade off.
Speaker 2 (42:03):
Yes, quality over quantity, yes, so I think that
the, the delay tool, is allowingyou to strike that balance
right, you can get in, you canscale it easily and get the
right volume, but yourconversion rates are still way
higher and you know I have.
It takes some weeks to get thiswhole thing off the ground
(42:24):
because you are shaping it right, it's not.
You have to tell it what youwant it to do and then you have
to like you're not just going tosend out these.
It produced an email or messageand you want to put it in
LinkedIn.
You've got to validate it.
I'm like, I'm good with that,or put in your own little
personal touch.
So it definitely improves theconversions and I think that's
(42:46):
why people are adopting this,because it's working.
And you know, if we wanted it,we just didn't have the
technology that was enabling usto get there.
Speaker 1 (42:55):
Well, actually, I would argue that we did have the
technology.
We just didn't have the human,like we didn't have the time to
take advantage of it the way wewould want to.
Speaker 2 (43:03):
That's fair, but the technology allows us to speed it
up.
Speaker 1 (43:08):
Yes, totally agree with that and I like this.
I guess this is what we'retalking about when I hear about
agents, the AI agents andagentic stuff.
That's what we're talking about.
Right, it's like being able todo multiple steps with a little
bit of guidance on what you wantthe output to be.
Speaker 2 (43:23):
And I mean I'm sure you've been there where it's
like you realize that the teamis spending their precious time
on these repetitive things andyou're like, let's outsource
that.
And then you're like, okay, nowwe got to find a human that's
going to do this outsourced,boring work, and then we got to
pay them extra and then we gotto check their work.
Speaker 1 (43:49):
Yeah, yeah.
So it's yeah, yeah, for sure.
Um, well, let's maybe one morething, um, and then we can.
We can wrap up so one of thethings I'm bullish on and have
been bullish on about potentialfor ai to have an impact on is
actually somebody we talkedabout earlier reporting and
analytics, because I do believeit has been a very it's like a
very human capital intensivething.
Right, there's a lot of effortin addition, just expertise, and
(44:10):
so I am bullish on that.
Be able to help really drivesome great insights over time
quickly, whether it's faster, uh, be able to do more um
experiments or test morehypotheses, whatever it is right
.
I think any of those would bebeneficial.
Are you seeing any of that alsoout there in your experience
(44:34):
with?
I know you're doing a lot ofconsulting these days, so what's
your take on that area of AI'simpact?
Speaker 2 (44:41):
Yeah, the biggest challenge with and I actually
had a conversation with a headof an agency today who does all
ops and digital marketing andshe literally said I am
constantly wreck, I'm spendingall my day reconciling data, so
(45:02):
I call it like it's alwaysstranded and siloed, so it lives
inside Salesforce and it livesinside HubSpot, and as much as
you're syncing these systems,you still have to make sure it's
right.
So time and time again I would,at the end of every week or
even month or quarter, I wouldpull data out of both my you
(45:23):
know run the business systemsand then I put it into a sheet
and then I create pivots and addcolumns, vlookups and all that
yep yeah, and then I share itwith the ceo and the cfo and
whoever else needs to know, andit's a pain in the butt.
And then you, you're, you're inthere and you're actually making
charts out of it and so on.
So AI your question is like howis AI helping?
(45:47):
That AI is the I referenced itearlier in our conversation
where it's like I've got thatlovely CSV file.
Like you can turn most of thatstructured data into CSV.
I put it into my AI system.
I chat in English, chart me X,y, z or you know how many.
How many deals did we closelast quarter that were higher
(46:10):
than 150K, example?
Right, and it will instantlyproduce the answer put it in a
table and pick your chart styleand then output and share that.
And the thing, the beautifulthing about having english chat
with your structured data is youknow you would automatically do
(46:31):
that in your meeting anyway.
Now you can actually talk tothe system to do it.
And a question begets anotherquestion.
I've been I say this all thetime so, as soon as you, one
question and you get a nicelittle nugget and then you're
like, well, hang on a minute,how about this?
And that's the beauty of thereally pulling wisdom out of
(46:55):
your data, right?
If you can read the data andask the right questions, then
you can really start to beinformed and knowledgeable.
So I'm all about dive in playwith it.
Speaker 1 (47:07):
ask more questions, ask more questions that
iterative process is how youlike, really get the most out of
it right.
But it takes again, takes timeand effort and expertise, and so
, again, if we can make thatfaster but that this is a
another aside, but like this iswhy I whole knowledge that these
things one question leads toanother question leads to
(47:28):
another is why, when I go tosome place and I go, we need to
get dashboards.
No, you don't, you really don'tneed dashboards.
No, start with something small,exactly If that becomes
something you know you're goingto do regularly, because you
usually don't know, fine, makethat the beginnings of a
dashboard, then do the next one,but don't know, fine, make that
the beginnings of a dashboard,then do the next one, but don't
start with the idea of we'regoing to do this comprehensive
(47:49):
dashboard.
Speaker 2 (47:50):
I think you think I've never seen one of those
succeed exactly, and if youpresent a dashboard which has
like 12 things on it that giveyou a little insight into each
part, then you you have to like,if you're talking to your CFO,
for example, which the boardwill actually ask as well, then
they're just glazing overbecause it's too much, yeah,
(48:12):
whereas like the nuggets like Ilove your point about like, just
start with something small andI can't tell you how many
meetings I've been in and youpresent what you believe they
want to hear.
You go to the meeting with yournice, pretty and you know charts
and then, of course, they askmore questions and then you're
like well, I have to get back toyou on that because I didn't
(48:34):
produce that report and it slowseverything down yeah and or, or
, or maybe worse.
Speaker 1 (48:41):
You know you've got a dashboard with multiple reports
that are already tough toconsume but somebody goes that
one's not right, right, yeah,and, and then that that reduces
the credibility of all the otherones that may be.
Quote right.
I mean, first off, saying it'sright or wrong is I think it's
just a misunderstanding aboutwhat you're actually looking at.
(49:03):
But yeah, that's a whole that'sa podcast yeah, okay, well now.
So now I feel like I have got alot of catching up to do on all
these ai tools and getting someof the mundane stuff out of my
life.
Um, this has been great.
Thank you for for sharing.
If, if folks want to connectwith you or learn more about
what you're doing, or find yourpod, your, your blogs or
(49:26):
whatever you're doing, what'sthe best way for them to do that
?
Speaker 2 (49:30):
They can go on LinkedIn and hopefully Deirdre
Mann.
Speaker 1 (49:34):
It's a little tricky to spell but we'll make sure we
put a link to it somewhere alongthe way.
Speaker 2 (49:39):
Yeah, I do have a website.
It's called marketing enginetwocom, the number two, because
my my belief is that usuallyyounger companies they try it
and then they're like, okay,let's do this properly, b2.
And then you know just beingreally modern about your
approach.
So it's always a challenge.
(49:59):
So you can check me out thereor find me on LinkedIn and
message me directly.
Speaker 1 (50:04):
Sounds good.
I appreciate it, and you are,because we traded messages today
on the marketingopscom Slack,so you're there too, right?
Speaker 2 (50:11):
Absolutely.
Speaker 1 (50:12):
All right perfect.
Speaker 2 (50:12):
Thank you for all you do there, because I get lots of
learning from that.
Speaker 1 (50:17):
That's great.
Well, we appreciate it.
Thank you again.
Thanks to all of our listenersout there for continuing to
support us and sharing yourideas and topics on guests and,
as always, if you want to be aguest or have a topic you want
to share with us to cover, feelfree to reach out to Naomi, Mike
or me, and we'd be glad tofollow up with that Till next
(50:37):
time.
Bye, everybody.
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