December 10, 2025 • 35 mins
Dave Boyce is a go-to-market–focused technology executive, author of Freemium, and longtime SaaS operator and board member. Over two decades leading and advising B2B companies, he’s helped founders navigate the shift from sales-led to product-led, from gut-driven to instrumented, and now from manual GTM to AI-assisted systems. Today, Dave teaches revenue architecture and PLG through Winning by Design’s Growth Institute, works with leadership teams on Bow Tie–based growth models, and is all-in on how AI will reshape GTM as a true engineered system.
Discussed in this episode
- Why classic sales & marketing playbooks haven’t caught up to how modern buyers actually buy.
- How the Bowtie model exposes the real levers of growth that funnels hide.
- Why PLG-style thinking is now essential even for sales-led and enterprise motions.
- The 3 first principles of freemium: empathy, generosity, and metrics.
- Where AI can reliably outperform humans across the customer journey, and where it absolutely shouldn’t.
- How to design hybrid human + AI workflows using a clear data model, not vibes.
- What RevOps should own in a modern revenue architecture (and why it can’t just serve the CRO narrative).
- Hard-earned founder lessons from Fundly on reinvention, calling bets early, and letting go of old branches.
Episode highlights
00:00 — GTM is still running 20-year-old playbooks
01:29 — “Sales, marketing, CS… the last unengineered engine”
03:20 — The myth of “just add more heads”
05:50 — The Fundly story: reinvention, too late
08:30 — Why Freemium had to be written
11:01 — Three first principles of freemium
15:25 — Mapping AI across the entire customer journey
19:29 — “Automate the predictable, humanize the exceptional”
25:18 — What the Bowtie exposes that funnels hide
27:25 — Building a “minimum viable Bowtie
This episode is brought to you by our sponsor: ZoomInfo
ZoomInfo is the GTM Intelligence Platform built for sales, marketing, and RevOps.
By unifying data, workflows, and insights into a single system, ZoomInfo helps revenue teams find and engage the right buyers, launch go-to-market plays faster, and drive predictable growth. With industry-leading accuracy and depth of data, it gives your team the intelligence advantage to win in competitive markets.
It’s trusted by the fastest-growing companies and has become the category leader in GTM Intelligence.
Learn more at zoominfo.com.
Follow Dave Boyce (Guest)
- LinkedIn: https://www.linkedin.com/in/boycedave
- Substack: https://daveboyce.substack.com
Where to Find GTMnow
- Website: https://gtmnow.com
- LinkedIn: https://www.linkedin.com/company/gtmnow
- X (Twitter): https://x.com/GTMnow_
- YouTube: https://www.youtube.com/@GTM_now
- Podcast Directory: https://gtmnow.com/tag/podcast
The GTMnow Podcast
The GTMnow Podcast is a weekly podcast featuring interviews with the top 1% GTM executives, VCs, and founders. Conversations reveal the unshared details behind how they have grown companies, and the go-to-market strategies responsible for shaping that growth.
Visit gtmnow.com for more episodes and other interesting content.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Dave Boyce (00:00):
GTM has not gotten that overhaul.
We're still doing what welearned to do 20 years ago, and
now we think it's still gonnawork today.
Sophie Buonassisi (00:06):
In this episode, Dave Boyce, who scaled
winning by design's revenuearchitecture practice, breaks
down the steps to rebuild thego-to-market engine for an AI
forward world.
He explains why PLG principlesnow sit at the center of
AI-driven revenue and theshifting buying landscape.
Dave Boyce (00:22):
Buyers are in control, sellers are not in
control.
Sophie Buonassisi (00:25):
We cover how to map your customer journey
into a measurable architecture,where AI can reliably outperform
humans, and what hybrid humanAI workflows actually look like
when well designed.
We also dig into the Bowtiemodel, which maps the entire
customer journey, both pre-saleand post-sale, in a single
connected system.
All right, let's get into it.
(00:56):
Dave, welcome to the podcast.
Dave Boyce (00:58):
Thank you so much, Sophie.
I can always tell when it's apro.
Sophie Buonassisi (01:02):
We are gonna have some fun.
And I have so many things toask you.
So let's dive right in.
Rock on because you set at thisreally interesting in
intersection where you couldjust see a lot of different
companies.
And so given the patterns thatyou're seeing across dozens of
go-to-market organizations rightnow, what part of the
go-to-market engine isstructurally breaking?
(01:24):
What's changing?
And what do leaders reallystill think is fine, but we
should be evolving towards?
Dave Boyce (01:29):
In my opinion, you know, talk about that as like
sales and marketing, and I wouldinclude customer success and
account management, is um is oneof the last kind of major
economic engines that hasn'tbeen engineered.
Sophie Buonassisi (01:46):
Yeah, and we hear all the time about
playbooks, right?
Some are becoming outdated.
And I know ourselves on theinvestor side, it's super
interesting because we'll helpfacilitate uh the connections
between a lot of go-to-marketexperts and portfolio companies.
And the playbooks that workedfor some operators no longer
apply to every single company.
They all require their ownunique nuance.
So I'm curious, like from yourperspective, also seeing a lot
(02:09):
of different companies, what isthat evidence that people are
still clinging to these kind ofoutdated models and not crossing
the chasm that you said?
Dave Boyce (02:18):
I mean, gosh, Sophie, like I mean, this is
gonna reveal my age, but youknow, I was taught um never demo
without discovery.
Like always figure, always dodiscovery first and then a demo.
That's one piece of kind ofpower that you have.
Never give a price withoutgetting access to someone who
(02:38):
controls a budget.
Um, those are things I wastaught.
That ain't true today.
Like buyers know how to findwhat they need, buyers know what
they need when they need it.
We're not in control, sellersare not in control, the
company's not in control.
But then what's the evidencethat we're missing that signal?
I mean, literally, like lastweek, a CEO came to us all
(03:00):
excited about engineering GTM,brought his brand new CMO and
brand new CRO with him.
They want to get from 100million to 200 million in ARR.
Okay, great, amazing.
How are you gonna do that?
We're gonna add 15% to ourheadcount, we're gonna spend
more on marketing, and we'regonna get and we're gonna move
up market.
Sophie Buonassisi (03:20):
Heard that before?
But AI is a big catalyst ofthat.
And you've been, I mean, deepin the trenches of AI.
We ourselves, we hear obviouslya lot of excitement around AI.
We see a lot of AI tools frommore of a startup investment
perspective, and we hear a lotof excitement across our LPs in
in the operator side of thehouse.
Everything from you know, emailwriters, meeting summarizers.
(03:43):
I think I've seen multiplemeeting summarizers just this
week alone.
But it seems like, and we bothknow the real shift is is far
deeper than better toolingindividually.
Like, what are the biggestmisconceptions anyone has about
AI's impact on revenue rightnow?
Like, how should we be thinkingabout this?
Dave Boyce (04:01):
We gotta start thinking about growth as a
system.
Sophie Buonassisi (04:04):
So if you had to actually describe the
current state of go-to-market inone sentence, let's say, what
would it be?
Dave Boyce (04:11):
Well, I'll give you an aspirational statement.
I think it's more designed,more architected, more
engineered, more closelymonitored, more iterated, uh,
more closely managed.
That's the operation, that'sthe aspirational kind of
statement.
Most of us are not there, butthe ones who are, like you look
at the kind of growthhyperscalers, they are that.
Sophie Buonassisi (04:32):
Yeah.
Dave Boyce (04:33):
They are that.
And many of them came out of aPLG background because PLG
requires you to be that, becauseyou have to program GTM into
the process, into the product,into the systems, because it's
going to run on its own.
Um, those of us who didn'tbuild PLG companies, you know,
we were able to get away withkind of like, and then she'll do
something great, and then he'llpick it up and he'll he'll
(04:55):
figure it out, and then youknow, I'll hire someone who
really knows how to do thatpiece of it.
And it's very kind of bespoke.
Um so aspirationally, we'remore designed, more engineered,
more calibrated.
But in reality, I think most ofus are on notice right now as
we watch these hyperscalers, andwe gotta retool.
Sophie Buonassisi (05:14):
Mm-hmm.
Yeah, it really has leveled theplaying field, is what it feels
like.
And I hear you on PLG companiesperhaps having an advantage
because they've had to designthese systems from the
beginning.
Yeah.
But you yourself, you know, youyou were a founder, you've
built a company, Fundly, andit's kind of one of those rare
founder journeys where the thestakes are existential as most
(05:36):
are.
So take us back to Fundly.
Love to hear a little bit moreabout what that reinvention
looked like for you.
And I know you had a crazy kindof time fundraising and
everything.
So love to hear a little bitmore on that story.
Dave Boyce (05:50):
So you want to pull me into my trauma space, huh?
Sophie Buonassisi (05:54):
Yes, this is actually a counseling session.
Okay, fantastic.
Dave Boyce (05:56):
Let me pull up a couch.
Um well, you know, it I thinkthe experiences I had as a
founder are not rare amongfounders, but it's but very few
of us step into that spacebecause it's scary.
And once we do, we realize whywe didn't.
Because it's super scary, it'ssuper hard.
Um, and you use the wordreinvention.
(06:18):
I I think um, you know what,funnily, we um I'll I'll say
this.
Did we reinvent?
Yes.
Did we get to uh good uniteconomics?
Yes.
Did we do it in time?
No.
Like, and it was because of me.
It's because I couldn't call itsoon enough.
I couldn't call balls andstrikes soon enough.
I couldn't tell someone thatthe thing that they had built
(06:41):
was gonna get disassembled andwe were gonna build it anew.
I mean, I did, but I just didit too slowly.
Like, this person who we gethired for that reason is now
fired, and now we're gonna pulla new person in.
Those are like human decisions.
The work that I did to buildkind of in that direction is not
working, so I'm gonnadisassemble it and we're gonna
go in a different direction.
(07:01):
The money that I raised fromthose investors based on what I
told them days ago or weeks ago,is now gonna be deployed in a
different direction.
But I'll tell you, Sophie, Ihad a I had a hard time with it,
which is why we did it toolate.
You know, we raised a seedround, it was super um super
encouraging.
(07:22):
Um we didn't pivot after theseed round.
We kind of like tried tostraddle strategies.
We'll keep this one goingbecause it's making us money.
While we launch our kind ofnext thing, we'll try to not let
go of the branch we're onbefore we have a firm grasp of
the next branch.
It moves it moves faster thatthan that.
I found a world.
(07:42):
You just freaking you get alittle money in the tank.
This one's not gonna be yourfuture, you just let go.
You drop it, you literally dropit, and you go to the next
thing, and that is super scary.
And when I did that after theseries A, that's when everything
started working.
Um, but we were a firm numbertwo in the market, and we should
have been number one.
Sophie Buonassisi (08:02):
Yeah, that's I mean, I'm sure you could write
a book on all the learningsfrom the director, funnily.
Dave Boyce (08:08):
What a good idea.
Sophie Buonassisi (08:10):
But you did just write a book, freemium.
Dave Boyce (08:12):
Yeah.
Sophie Buonassisi (08:13):
And congratulations for some
premiums.
It is perfectly timed, as we'vediscussed, with PLG
intersecting with AI, and asgo-to-market teams really
rethink cost structure too.
What was the spark that madeyou decide that this book needed
to exist?
Dave Boyce (08:30):
Oh man, I you know, it is perfectly timed, but I
didn't think so.
Like I we got when I so I saw along-term trend.
Um I believed that self-servicebuying is a long-term trend.
It started kind of in thedigital era with e-commerce, and
then it came to then e-commercekind of pushed onto our mobile
phones, and then and then SaaSbecame a thing, and then SaaS
(08:53):
became self-service through PLG,and then PLG, you know,
extended into more and morecomplex use cases through AI
assistive kind of buying.
I could see that trend.
I could see it coming, and Iknew that the definitive book
had never been written, eventhough we had companies like
Atlassian and Canva and Twilioand DocuSign and Dropbox that
had built amazing companiesbased on self-service kind of
(09:15):
PLG models.
No one had written thedefinitive book, and I was like,
can that really be true?
Like I called my friend MarkRoberge and I'm like, I'm I'm
teaching this now at uh in theMBA program at BYU.
I'm gonna write a book.
He's like, okay, great, youshould write the book.
It needs to be written forpeople who didn't build that
way, not people who did buildthat way, because we had a lot
(09:36):
of kind of blogosphere stuffgoing around PLG.
You almost couldn't turn aroundat Silicon Valley without
hearing PLG.
This is three years ago, so ifyou like, and um he's like, and
I'm like, okay, cool.
So which cases should I use?
And he's like, Yeah, therearen't really any cases out
there.
I'm like, what?
Like, I go into HarvardBusiness Publishing and I just
search on PLG and I come upempty.
(09:57):
He's like, Yeah, because it isempty.
Like, here's a case on Dropboxthat you can kind of repurpose.
Here's a case that I've writtenthat has that's not in there
yet.
And other than that, you're onyour own.
This is three years ago.
So I'm like, it's crazy becausewe actually know how to do
this, but no one's written thebook.
So that was the spark.
Sophie Buonassisi (10:17):
Incredible.
I'm glad it worked out.
Dave Boyce (10:18):
Yeah.
Sophie Buonassisi (10:19):
And glad we can talk about it today.
No, we're big believers here atGTM Fun that strategy alone
doesn't win deals, executiondoes.
But 85% of sellers are stuckmanaging their books in
spreadsheets because theirdata's scattered across dozens
of tools.
ZoomInfo Copilot Workspacefixes that.
It's the first workspace wheresellers actually work.
Complete buyer context, yourREM data, and AI-powered
(10:40):
insights in one place.
Execution wins deals,everything else is really just
preparation.
You can learn more atzoominfo.com forward slash
copilot.
It'll be in the show notes.
What are the most foundationalparts of it that you think if
you could tell any go-to-marketleader or founder listening,
hey, remember these three thingsor two things?
Dave Boyce (11:01):
Yeah.
Well, for the first thing, justto get your attention, I'd just
ask you like, when's the lasttime you spoke to a human to buy
gas for your car?
Sophie Buonassisi (11:10):
I try to avoid it at all costs and all
that.
Dave Boyce (11:12):
Of course.
Like, why would like why ofcourse?
And then you think, well, whatabout all the other stuff in my
life?
When is the last time you spoketo a human to buy groceries?
When's the last time you spoketo a human to buy an app on your
phone?
When's the last time you spoketo a human to buy a piece of
software that you use in yourwork life?
Like, it is less and lessfrequent over time, less and
less frequent.
(11:33):
That's just a long-term trend.
So then you think, all right,well, my competitors, or sorry,
my customers are versions of me.
So how much do they want totalk to me?
How much do they want toschedule a demo?
How much do they want to get onlike get on a phone call, a
discovery call?
Like zero.
They don't either.
(11:53):
And if they and if if they canfigure out a way to kind of
trial their way into successwith my competitor's product, or
just get their hands on mycompetitor's product where I'm
putting up walls, they're gonnago with my competitor.
That's just the way it works.
So then so if if we can kind oflike break out of this notion
(12:16):
that it matters at all how we'reused to doing it, because it
doesn't.
The market cares zero.zero howwe're used to doing it.
The market just wants what whatit wants and it's gonna get it
wherever it can get it.
If it's from me, great.
If it's from my competitors,great.
Market doesn't care.
So here we are kind of holdingon to the old ways.
Here's our competitor saying,Oh, you want it that way?
Cool, here, try it.
(12:37):
Then you just got to thinkabout what are the first
principles that would help youbuild in that way.
And now I'm asking, finallyanswering your question.
These are three firstprinciples in the book, and then
we go into all the details forhow you engineer this into your
product.
But the first is empathy, andthat's and unironically,
(12:57):
empathy, like literally, I wantto understand the end user of my
product.
I want to understand how shedefines progress in her life,
how she what her jobs to be doneare, and where and when it
would make sense for her to hiremy product or service to help
her make progress.
I really want to understandthat what are the words she
uses, what are them, how doesshe measure it, how does she
experience it.
Okay, cool.
(13:18):
So now once I understand that,I want to build something that's
going to kind of meet her whereshe is.
I'm not going to require her tomeet me where I am, I'm going
to meet her where she is.
And the next principle, again,unironically, is generosity.
Oh, that's the problem?
That's what you're trying.
Here, try this.
Cool, what do I owe you?
Nothing.
(13:38):
When do I pay you?
Don't worry about it.
Never.
Like, let's just see if itworks.
Like, and then once she'srolling and she's and she's
developing habit around that,eventually I will, of course.
She'll get to a point where shewants to do a longer than a
40-minute Zoom call or scorestore more than 90 days of Slag
messages or or two terabytes ofstorage or whatever.
(14:00):
She'll get to a point, shewants one access to templates or
you know, whatever.
Great.
And then she'll be happy topay.
So one is empathy, two isgenerosity.
And then the third thing is ifwe're building our products this
way, we need metrics.
Because we don't have a humanin the room here observing how
she's experiencing the product.
We don't have a human in theroom kind of coaching her to do
(14:22):
the next thing, training her,um, teaching her.
So we're relying on the productto do that work, and the only
way we know if it's working ornot is by metrics.
Like we will see the clicktracks, we will see the success
paths, we will see the deadends, we'll see where where
she's not discovering somethingwe thought she should have
discovered, and we'll see thatall in the metrics and in the
(14:43):
analytics, and then that's howwe will fine-tune our product
over time.
So empathy, generosity,metrics.
There's a whole bunch ofscience behind how you do that
and cohorted-based kind ofmeasurement, and but if those
are the three first principlesI've got to get into place if I
want to meet today's buyerswhere they live.
Sophie Buonassisi (15:00):
Mm-hmm.
Yeah, I love the simplicity andit it sounds simple the way that
you framed it, however, it issuper hard in practice.
Dave Boyce (15:07):
Super hard, yeah.
Sophie Buonassisi (15:09):
Yeah.
And how does AI come into playwith the book?
Because if we map the buyerjourney from end to end, AI is
touching more and more of it.
What is AI's impact on go tomarket right now?
Like, where is it most visiblethat you're seeing and that
you've written into the book?
Dave Boyce (15:25):
If I think about that customer journey, and in
you know, PLG terms, we're gonnatalk about um we're gonna talk
about awareness, we're gonnatalk about acquisition,
activation, first impact, thenwe're gonna talk about habit,
then monetization, thenretention, then engagement, then
(15:49):
retention, then expansion.
So these are kind of likestages in the customer journey.
Okay, cool.
So we've mapped that all out.
And then you think, all right,well, where could AI help?
It can help almost anywhere.
Like, you know, I'm let's sayI'm active, I've created my
account and now I want toactivate.
AI can can be almost like aguide to help me do the things.
It can make guesses for me.
(16:10):
I don't know if you've everused any kind of vibe coding
platforms or like an AIautomatic website generator.
You tell it a few prompts andit just says, here, let me
guess.
Here's some pictures, here's alogo, here's a um, here's a uh
login kind of um I want to saydialogue.
Um it just pulls it out of itslibrary, pulls it out of its
(16:31):
memory, does pattern recognitionand kind of fills in the gaps.
Cool, that's a way fasteractivation process than if I had
had to kind of read somethingand go do it myself.
You can think about it foronboarding.
You think about it for um forfeature uh discovery, like hey,
I'm I'm chugging along in Canva,and Canva identifies that I may
(16:54):
be able to benefit fromtemplates, or I may be able to
benefit from a backgrounderaser, or I might be able to
benefit from something that Ihaven't discovered yet.
Boom, it can suggest it for me.
Um and then you can think aboutlong-term engagement, like it
just helps me, AI just helps meum accomplish what I want to
accomplish.
Uh helps me know when I shouldupgrade to the next tier, it
(17:16):
helps me know when I shouldexpand this to team usage, etc.
etc.
The other thing that AI, sothat's just basic.
That's basic, Sophie, justbasic kind of automating the
predictable of all the wayacross the customer journey.
The other thing that AI can dothough is unlock product
categories that were previouslynot automatable.
So now I have a very complexkind of, you know, think about
(17:38):
the most complex software thereis, you know, like uh process
manufacturing ERP.
Like, oh my gosh, like supercomplex.
Um, lots of configuration, lotsof connections to uh machines,
etc.
Well, AI can kind ofdecomplexify that for me too,
like literally tutorial kind ofhelp me connect this machine,
(17:59):
help me configure this uhdashboard, help me interpret
these metrics.
And AI can literally just helpme self-serve my way into
success where I may have neededtwo or three humans and two or
three months to do it in thepast, and now a previously too
complex to be self-serviceproduct becomes self-service.
Sophie Buonassisi (18:20):
Which is incredible.
And we're seeing it happen moreand more.
I mean, you go to these Yarmywebsites, even you can see these
product led growth, these PLGflows in place.
Dave Boyce (18:30):
I mean, even S I I was just in Saudi Arabia working
with a manufacturing firm,they're doing that same thing.
Even SAP, you know, largelythought of as the most kind of
most monolithic and complicatedsoftware, they've got agents all
throughout their go-to-market,and they've got like almost a
PLG kind of flow where you firststart in a sandbox that has
their data, then you go to anoffline sandbox where you've
(18:54):
uploaded your uploaded your owndata, then you go to like a
cloud environment where you canbe more real-time, and then you
go to um like the fully kind ofpermissioned you know, SAP
instance.
But that's like a self-serviceonboarding that you would never
have thought of five years agofor SAP.
Sophie Buonassisi (19:14):
Yeah, exactly.
So what what tasks should AIimmediately own, and which ones
should humans protect?
Just it sounds like we'removing further and further along
the spectrum, which we know weare, but how should people be
thinking about that divide?
Dave Boyce (19:29):
I love it.
There's a um there's gonna be aweird way to start this answer,
Sophie, but there's a a Polishfantasy fiction author named
Joanna.
She says, I want AI to do mydishes and laundry so that I can
do art and writing.
I don't want AI to do art andwriting so that I have to do
dishes and laundry.
I really think that is how weshould think about it.
(19:51):
We want AI to automate thepredictable so that we we can
humanize the exceptional.
So anything that's Predictable.
Fill out a form, fill out afield, process an order, build a
do research on an account, youknow, respond to a trigger with
(20:12):
uh with some sort of a you knowjust a mechanical kind of
acknowledgement.
All of that should beautomated.
AI should be doing all of thatfor us.
And then the human stuff, ifyou're thinking in terms of go
to market, the the human, it'swhere we want to show up as a
real human, like the emotionalstuff, like many times on the
(20:33):
other side is a is a real human.
This is me and you talking,like if I'm the if I'm let's say
I'm with the selling company orwith the buying company, you're
trying to accomplish something.
Like you're not just out herefor fun and you've gotten a
certain way amount of the way onyour own, and now you're
trying, now you're trying tohave the courage to put this
(20:54):
forward as the standard platformwithin your company.
Okay, courage is something thatwe can work on together.
I can help you connect withpeople that have done the same
thing.
I can use my judgment or mypattern recognition and and then
and also my trust building withyou and and help you get to the
point where you're managingyour stakeholders in a way that
(21:15):
makes sense for kind of whatthey're gonna need in order to
approve that as a go forwardthing.
That's a very human experience.
It would be very hard for youto trust a robot who was kind of
trying to coach you on stufflike that.
But when we get to the pointwhere all the automated stuff,
all the predictable stuff isautomated, then that means we
have to show up as likesuperhumans.
(21:37):
Like we gotta show up likereally, really human, like
aesthetic and intuitive andhelpful, and in a how can I help
you mode, not a what can I sellyou mode.
And that's that's the versionof the future that I want to
believe in.
Well, the first thing you need,um, you need a theory of the
(21:58):
case, Sophie.
Like, um, and I think that allstarts with the data model.
Like we described the PLG datamodel from you know, awareness
acquisition, activation, firstimpact habit, you know, blah,
blah, blah.
I need a theory of the case.
I need to know what that lookslike.
I need to and I I need to havekind of mapped it out into what
I would call a data model.
Winning by design uses thebowtie data model.
(22:20):
That's the PLG version of it.
The sales-led version of itwould be um awareness,
education, selection, commit,and that's like the traditional
sales and marketing funnel, kindof narrowing as you go.
Awareness, education,selection, commit.
Commit is the narrowest part ofthe funnel.
I have it turned on its side,but it's still a funnel.
(22:41):
And then that's we that's likethe knot of the bow tie, and
then I'm gonna start opening outfrom there.
Onboarding, retention, andexpansion is gonna hopefully
kind of make my long-termlifetime value with that
customer actually expand overtime as we deliver impact.
Okay, now if I can define eachof those um stages and I can
(23:02):
define what needs to happenwithin each of those stages, you
can imagine that that is like aum almost like a manufacturing
process, right?
Something happens during thisstage, something happens during
this stage, there's a handoff,something happens during this
stage.
There's success criteria,there's activities in stage,
there's success criteria to exitthe stage.
And then once I have it definedlike that, that's kind of like
(23:23):
bare minimum, and it's not alot.
Like that's bare minimum for meto be able to start running
experiments.
Then I can start A-B testinghuman versus robot on this task.
Did it help me improve myconversion rate?
Oh, it didn't?
Okay, then that's probably notthe right place to be using a
robot.
How about or maybe I tune it alittle more and see if I can get
it there, tune it a littlemore, eventually I get it there.
Oh, cool, cool.
(23:43):
The robot can help me there.
It helps make my reps moreefficient, it helps make my
conversion rates better, but Igotta be able to measure it,
which means I need a data model.
And that data model needs tonot be a political data model.
I'm sure you've been in boardmeetings, I've been in uh board
meetings where it feels like I'mjust on the receiving end of
like a commercial and theoperators are just trying to
(24:04):
kind of convince me that nothingstinks in this business.
That ain't that's not what weneed.
We need a very clear light ofday, consistent data model so
that we can run experiments,because we're just gonna run
experiments and we're gonna getAI deployed everywhere that it
works, and if we deploy it andit doesn't work, no harm, no
foul, we'll kick it out andwe'll try it somewhere else.
So I think data model first,kind of theory of the case.
(24:26):
What are we trying toaccomplish?
And then we can start choosingwhere we're gonna deploy AI, and
we'll treat it as an experimentuntil it's proven that it's a
permanent thing.
In lean manufacturing, this islike one of my favorite things
about uh sayings in leanmanufacturing that you never
hear.
You always hear aboutcontinuous improvement or the
and-on cord or uh just in time,but here's here's a very cool
(24:47):
one.
Don't bolt down what you can'ttape down, and don't tape down
what you can't hold down.
So basically, I don't need togo bolting the AI in place when
I still don't know if it's gonnawork.
Like, let me just hold it inplace.
Let me test it.
Oh, that looks like it's gonnawork.
Now let me tape it and let mestep away for a second.
(25:09):
Oh, it still looks like it'sholding.
Okay, now we're convinced AIcan do that job consistently,
reliably, cool.
We're gonna bolt it down andmove on to the next experiment.
Sophie Buonassisi (25:18):
What does the bow tie model actually expose
that traditional funnels hidewhen winning by design is
employing it?
Dave Boyce (25:24):
That's where the renewals happen, that's where
the expansion happens, andthat's where the growth loops
are initiated that can pull newcustomers into the front of the
funnel based on referrals fromexisting customers.
All the compounding happens onthe right hand side of the bow
tie.
Now we spent, if we if you grewup when I did or anytime, you
know, if you built a companyanytime before the last five
(25:47):
years, you might have spent aton of time in QBRs and planning
sessions and board meetingstalking about calling and
meeting your bookings forecast.
That's basically what WallStreet tracks.
That's basically what you knowmost sales-led organizations
track.
It's where we put our mostexpensive people, it's where we
(26:09):
put all of our executiveattention, it's where we put all
of our focus in those meetings,is calling and hitting a
bookings forecast.
But bookings is the knot of thebow tie.
It's the beginning of thejourney.
Everything after that is thecustomer's experience, and it if
that customer experience isgood, then renewals and
expansion will happen, whichmeans now I have the machine
working for me instead of meworking the other way around.
(26:31):
So why wouldn't I also bespending time and attention
there?
Why do I put all of myexpensive time and resources and
people and attention onbringing new customers in and I
short sheet the right-hand sideof the bow tie?
It just doesn't make sense froma systems or math perspective.
And uh and the bow tie kind ofjust like brings that to life
because once you start runningcohorted math through that
(26:53):
system, you start seeing like,oh my gosh, this short term,
yes, short term, I will get alot of benefit from sales, but
long term compounding growth isall driven on the right side of
the bow tie.
Sophie Buonassisi (27:07):
Interesting with those cohorted systems are
real.
So, what would be if we take uhan MVP, for example, for
inspiration on the product side,what's like a minimum viable
bow tie for a company to need tobe able to run more of a hybrid
AI and uh PLG or go-to-marketmotion?
Dave Boyce (27:25):
Yeah.
The hardest thing is connectingthe right and left side of the
bow tie.
So all almost all of us havethe left side of the bow tie
instrument in some way, shape,or form.
We have stage one, stage two,stage three opportunities.
We've got it all in HubSpot orSalesforce.
We kind of we we kind of knowhow to build our kind of
bookings forecast and we manageit, and we have MQLs and SQLs,
like we have that built.
Sophie Buonassisi (27:45):
Yeah.
Dave Boyce (27:46):
Where does the stuff on the right hand side live?
Sometimes it lives in somethinglike a gain site, sometimes
we've written it back to ourCRM, and sometimes it's neither
place, and you have to go get itout of finance.
It's literally tracked based onbillings because we don't have
it in CRM and we're not trackingit in a CS platform, and we
literally have to say, well,when did we, you know, did we
(28:08):
send that customer a bill or notbased on whether they canceled
or not?
So minimum viable product wouldbe getting the left hand and
the right hand side stitchedtogether all the way through the
journey that I can see in oneplace.
Once I have that, and if I'vedone it according to the winning
by design uh bow tie dataschema, then I can benchmark it.
(28:28):
We've got 300 companiesbenchmarked, and you can cohort
that based on companies that aresimilar to you or on similar
motions to you, and then you canstart benchmarking.
But even if you can'tbenchmark, at least you can
compare yourself to last periodon the period before.
Now you've got like a baselineand you can start seeing if if
you're actually makingimprovements.
But before you have that, Idon't think you can improve a
(28:50):
human system.
And I also don't in a reliable,consistent, kind of ongoing way
and like a continuousimprovement in lean
manufacturing way, and youcertainly can't improve an AI
system or human AI hybrid systembecause you just don't have the
instrumentation to tell umwhat's working and what's not.
Sophie Buonassisi (29:07):
So when teams are wanting to create this bow
tie framework and they need toeither take their data from
GainSite or some other system,where are they creating the bow
tie?
Where's this connectivitybetween the left side and the
right side of covering?
Dave Boyce (29:22):
So ideally you would write everything back to CRM.
That that's e way easier saidthan done.
CRM wasn't necessarily builtfor things like growth loops.
Um very tough to do that.
Certainly not built for thingslike um you know activation of
an account pre um pre-payment.
(29:43):
Um that's a PLG thing, likevery difficult.
But let's say you could get itinto CRM.
That would be my choice A.
Um, but what we very often seeis we'll get it into like
Snowflake.
Um or we'll get it intoSnowflake and then a
visualization layer like a umlike a Domo or uh uh you know
(30:05):
some sort of or Power BI or somekind of visualization layer.
There are also some uh productsout there that are doing this
commercially um that arepartners of Winning by Design
that that will take that willvisualize the bowtie for you,
like a UNA or a Vasco.
So like SaaStrack is a partnerof Winning by Design, adheres to
the bowtie data schema, buildsthe custom objects, it's a it's
(30:26):
a managed package, builds thecustom objects inside of
Salesforce for you, and now youcan just kind of manage it and
use Salesforce Reporting.
Amazing.
That's a SaaSTrek.
Vasco and Una pull it all out,like if you have it in disparate
systems and they give youvisualization and a management
framework.
Or you can do it like I said,roll your own and put it in your
own BI tool.
But you do want that wiring tobe um you want it to be wired,
(30:51):
not just kind of like CSV kindof one-time pulls.
Because you track it thismonth.
The continuous exactly.
Yeah.
Sophie Buonassisi (31:00):
Right, right.
Super, super.
Who owns that process?
Dave Boyce (31:04):
Jeez, you are so mean.
You're asking all the hardestquestions.
Ums, it's not very consistent.
Ideally, you would have afunction called RevOps that
didn't just work for sales.
And ideally, that RevOpsfunction would build those
(31:25):
systems and would be the kind ofimpartial, objective arbiter of
truth.
In many companies, that's notwhat RevOps is.
In many companies, the RevOpsdoes whatever it takes to make
the CRO look good.
And they help them prepare forboard meetings, and they help
them kind of scrub and whitewashnumbers, and they help them
kind of um maybe maybe uh inprivate they're look looking for
(31:50):
truth, but in public they'rewhitewashing.
That ain't gonna work.
That's just not gonna work.
So ideally we just we just stepRevOps up into a kind of
impartial arbiter of truth, andif not, then it can also be
FPNA.
But the problem with FPNA isthey don't actually understand
the go-to-market well enough andthe systems that run
go-to-market well enough.
So my ideal, Sophie, would bethat RevOps steps into this role
(32:14):
going forward and is thebackbone of our modern
go-to-market.
Sophie Buonassisi (32:19):
I love it.
And for other operators, notjust RevOps, I have heard you
say that.
They need to become more chieffigure it out officers, which I
think is a really fun term.
What does that role look likein the day-to-day?
What does that entail?
Dave Boyce (32:35):
Full attribution.
That's Ryan Sanders fromMercado.
He gave me that term, and I'vebeen I've been shamelessly
reusing it.
Chief figure it out officer,it's super easy to remember.
But it's all the stuff that youand I have been talking about,
like where would I deploy AI?
I don't know.
I can't pull that out of my bagof tricks.
I can't go, you know, rewindthe clock 15 years to when I
(32:58):
used to be an AE and tell youhow I used to use AI.
I didn't use AI, which means wegot to figure this out
together.
I certainly didn't use itsystematically.
I can't tell my RevOps personwhat we did 15 years ago to
instrument or go to market sothat we could run A-B
experiments with human robothybrid systems because we didn't
have that.
(33:18):
So we got to figure all of thisout as we go, which means if
I'm if I'm a veteran, which Iam, um if I'm a veteran head of
revenue, I gotta get out of theidea that I'm just gonna teach
people how to how to do it theway I did it.
And I gotta get into themindset of no, no, no, we're
(33:38):
gonna figure this out together.
So I'm gonna grab smart peopleon my left and right, we're
gonna go in, we're gonnasystematically architect
something that I can then run byinstrumentation versus running
in an Amelia Earhart way, like,ooh, it looks cloudy over there,
I better steer left.
Sophie Buonassisi (33:54):
So you've obviously spent a lot of time
writing your own book.
Are there other booksthroughout your career that have
made a particular impact onyou?
Dave Boyce (34:00):
Oh, good question.
Yeah, I've I my my author herois Clayton Christensen.
Um he's amazing.
I've read, I think everythinghe's written.
I think he is he's a reallygood scholar.
Uh may you rest in peace.
Um I went to both of hisfunerals.
He was a good friend too, but Iknew him as a scholar and a
friend.
And um, so what I alwaysrecommend if you haven't read
(34:23):
anything of his is CompetingAgainst Luck.
It's not his most famous book,but it's a really good book for
this moment.
Um and that's where he reallyunveils the jobs to be done
theory.
Uh I super like um RogerMartin's book called Playing to
Win.
Um very, very good strategyframework.
(34:43):
Um I could go on for for daysabout books that have made an
impact, but that's where usuallywhere I start.
Sophie Buonassisi (34:52):
Very cool.
Great recommendations.
And what about yourself?
Where can people follow alongyour journey?
Obviously, we'll have a linkfor the book, but for yourself
at all, are you on LinkedIn X?
What's the best place to followyou?
Dave Boyce (35:02):
I'm not active on X.
I would love for obviously youcan find me here, and we're
gonna put that in the um shownotes.
I'm on LinkedIn, pretty activeon LinkedIn.
Uh I have a Substack uh whichis just Dave Boyce.
Um and uh and if you join theGrowth Institute with Winning by
Design, you'll see me superactive there.
I run quarterly case studies,MBA style, kind of executive
(35:26):
education case studies.
You'll see me on stage at thesummits, and we will uh and
we'll go change the worldtogether.
Sophie Buonassisi (35:34):
Amazing.
Amazing.
Dave, this has been phenomenal.
Thank you for the time.
Thank you for the book on mybookshelf.
Dave Boyce (35:40):
Yay!
Sophie Buonassisi (35:40):
And uh yeah, really, really appreciate it.
Dave Boyce (35:43):
Thanks so much, Sophie.
Amazing.
Let's go do it.
Sophie Buonassisi (35:46):
Let's do it.
GTM has not gotten that overhaul.
We're still doing what welearned to do 20 years ago, and
now we think it's still gonnawork today.
Sophie Buonassisi (00:06):
In this episode, Dave Boyce, who scaled
winning by design's revenuearchitecture practice, breaks
down the steps to rebuild thego-to-market engine for an AI
forward world.
He explains why PLG principlesnow sit at the center of
AI-driven revenue and theshifting buying landscape.
Dave Boyce (00:22):
Buyers are in control, sellers are not in
control.
Sophie Buonassisi (00:25):
We cover how to map your customer journey
into a measurable architecture,where AI can reliably outperform
humans, and what hybrid humanAI workflows actually look like
when well designed.
We also dig into the Bowtiemodel, which maps the entire
customer journey, both pre-saleand post-sale, in a single
connected system.
All right, let's get into it.
(00:56):
Dave, welcome to the podcast.
Dave Boyce (00:58):
Thank you so much, Sophie.
I can always tell when it's apro.
Sophie Buonassisi (01:02):
We are gonna have some fun.
And I have so many things toask you.
So let's dive right in.
Rock on because you set at thisreally interesting in
intersection where you couldjust see a lot of different
companies.
And so given the patterns thatyou're seeing across dozens of
go-to-market organizations rightnow, what part of the
go-to-market engine isstructurally breaking?
(01:24):
What's changing?
And what do leaders reallystill think is fine, but we
should be evolving towards?
Dave Boyce (01:29):
In my opinion, you know, talk about that as like
sales and marketing, and I wouldinclude customer success and
account management, is um is oneof the last kind of major
economic engines that hasn'tbeen engineered.
Sophie Buonassisi (01:46):
Yeah, and we hear all the time about
playbooks, right?
Some are becoming outdated.
And I know ourselves on theinvestor side, it's super
interesting because we'll helpfacilitate uh the connections
between a lot of go-to-marketexperts and portfolio companies.
And the playbooks that workedfor some operators no longer
apply to every single company.
They all require their ownunique nuance.
So I'm curious, like from yourperspective, also seeing a lot
(02:09):
of different companies, what isthat evidence that people are
still clinging to these kind ofoutdated models and not crossing
the chasm that you said?
Dave Boyce (02:18):
I mean, gosh, Sophie, like I mean, this is
gonna reveal my age, but youknow, I was taught um never demo
without discovery.
Like always figure, always dodiscovery first and then a demo.
That's one piece of kind ofpower that you have.
Never give a price withoutgetting access to someone who
(02:38):
controls a budget.
Um, those are things I wastaught.
That ain't true today.
Like buyers know how to findwhat they need, buyers know what
they need when they need it.
We're not in control, sellersare not in control, the
company's not in control.
But then what's the evidencethat we're missing that signal?
I mean, literally, like lastweek, a CEO came to us all
(03:00):
excited about engineering GTM,brought his brand new CMO and
brand new CRO with him.
They want to get from 100million to 200 million in ARR.
Okay, great, amazing.
How are you gonna do that?
We're gonna add 15% to ourheadcount, we're gonna spend
more on marketing, and we'regonna get and we're gonna move
up market.
Sophie Buonassisi (03:20):
Heard that before?
But AI is a big catalyst ofthat.
And you've been, I mean, deepin the trenches of AI.
We ourselves, we hear obviouslya lot of excitement around AI.
We see a lot of AI tools frommore of a startup investment
perspective, and we hear a lotof excitement across our LPs in
in the operator side of thehouse.
Everything from you know, emailwriters, meeting summarizers.
(03:43):
I think I've seen multiplemeeting summarizers just this
week alone.
But it seems like, and we bothknow the real shift is is far
deeper than better toolingindividually.
Like, what are the biggestmisconceptions anyone has about
AI's impact on revenue rightnow?
Like, how should we be thinkingabout this?
Dave Boyce (04:01):
We gotta start thinking about growth as a
system.
Sophie Buonassisi (04:04):
So if you had to actually describe the
current state of go-to-market inone sentence, let's say, what
would it be?
Dave Boyce (04:11):
Well, I'll give you an aspirational statement.
I think it's more designed,more architected, more
engineered, more closelymonitored, more iterated, uh,
more closely managed.
That's the operation, that'sthe aspirational kind of
statement.
Most of us are not there, butthe ones who are, like you look
at the kind of growthhyperscalers, they are that.
Sophie Buonassisi (04:32):
Yeah.
Dave Boyce (04:33):
They are that.
And many of them came out of aPLG background because PLG
requires you to be that, becauseyou have to program GTM into
the process, into the product,into the systems, because it's
going to run on its own.
Um, those of us who didn'tbuild PLG companies, you know,
we were able to get away withkind of like, and then she'll do
something great, and then he'llpick it up and he'll he'll
(04:55):
figure it out, and then youknow, I'll hire someone who
really knows how to do thatpiece of it.
And it's very kind of bespoke.
Um so aspirationally, we'remore designed, more engineered,
more calibrated.
But in reality, I think most ofus are on notice right now as
we watch these hyperscalers, andwe gotta retool.
Sophie Buonassisi (05:14):
Mm-hmm.
Yeah, it really has leveled theplaying field, is what it feels
like.
And I hear you on PLG companiesperhaps having an advantage
because they've had to designthese systems from the
beginning.
Yeah.
But you yourself, you know, youyou were a founder, you've
built a company, Fundly, andit's kind of one of those rare
founder journeys where the thestakes are existential as most
(05:36):
are.
So take us back to Fundly.
Love to hear a little bit moreabout what that reinvention
looked like for you.
And I know you had a crazy kindof time fundraising and
everything.
So love to hear a little bitmore on that story.
Dave Boyce (05:50):
So you want to pull me into my trauma space, huh?
Sophie Buonassisi (05:54):
Yes, this is actually a counseling session.
Okay, fantastic.
Dave Boyce (05:56):
Let me pull up a couch.
Um well, you know, it I thinkthe experiences I had as a
founder are not rare amongfounders, but it's but very few
of us step into that spacebecause it's scary.
And once we do, we realize whywe didn't.
Because it's super scary, it'ssuper hard.
Um, and you use the wordreinvention.
(06:18):
I I think um, you know what,funnily, we um I'll I'll say
this.
Did we reinvent?
Yes.
Did we get to uh good uniteconomics?
Yes.
Did we do it in time?
No.
Like, and it was because of me.
It's because I couldn't call itsoon enough.
I couldn't call balls andstrikes soon enough.
I couldn't tell someone thatthe thing that they had built
(06:41):
was gonna get disassembled andwe were gonna build it anew.
I mean, I did, but I just didit too slowly.
Like, this person who we gethired for that reason is now
fired, and now we're gonna pulla new person in.
Those are like human decisions.
The work that I did to buildkind of in that direction is not
working, so I'm gonnadisassemble it and we're gonna
go in a different direction.
(07:01):
The money that I raised fromthose investors based on what I
told them days ago or weeks ago,is now gonna be deployed in a
different direction.
But I'll tell you, Sophie, Ihad a I had a hard time with it,
which is why we did it toolate.
You know, we raised a seedround, it was super um super
encouraging.
(07:22):
Um we didn't pivot after theseed round.
We kind of like tried tostraddle strategies.
We'll keep this one goingbecause it's making us money.
While we launch our kind ofnext thing, we'll try to not let
go of the branch we're onbefore we have a firm grasp of
the next branch.
It moves it moves faster thatthan that.
I found a world.
(07:42):
You just freaking you get alittle money in the tank.
This one's not gonna be yourfuture, you just let go.
You drop it, you literally dropit, and you go to the next
thing, and that is super scary.
And when I did that after theseries A, that's when everything
started working.
Um, but we were a firm numbertwo in the market, and we should
have been number one.
Sophie Buonassisi (08:02):
Yeah, that's I mean, I'm sure you could write
a book on all the learningsfrom the director, funnily.
Dave Boyce (08:08):
What a good idea.
Sophie Buonassisi (08:10):
But you did just write a book, freemium.
Dave Boyce (08:12):
Yeah.
Sophie Buonassisi (08:13):
And congratulations for some
premiums.
It is perfectly timed, as we'vediscussed, with PLG
intersecting with AI, and asgo-to-market teams really
rethink cost structure too.
What was the spark that madeyou decide that this book needed
to exist?
Dave Boyce (08:30):
Oh man, I you know, it is perfectly timed, but I
didn't think so.
Like I we got when I so I saw along-term trend.
Um I believed that self-servicebuying is a long-term trend.
It started kind of in thedigital era with e-commerce, and
then it came to then e-commercekind of pushed onto our mobile
phones, and then and then SaaSbecame a thing, and then SaaS
(08:53):
became self-service through PLG,and then PLG, you know,
extended into more and morecomplex use cases through AI
assistive kind of buying.
I could see that trend.
I could see it coming, and Iknew that the definitive book
had never been written, eventhough we had companies like
Atlassian and Canva and Twilioand DocuSign and Dropbox that
had built amazing companiesbased on self-service kind of
(09:15):
PLG models.
No one had written thedefinitive book, and I was like,
can that really be true?
Like I called my friend MarkRoberge and I'm like, I'm I'm
teaching this now at uh in theMBA program at BYU.
I'm gonna write a book.
He's like, okay, great, youshould write the book.
It needs to be written forpeople who didn't build that
way, not people who did buildthat way, because we had a lot
(09:36):
of kind of blogosphere stuffgoing around PLG.
You almost couldn't turn aroundat Silicon Valley without
hearing PLG.
This is three years ago, so ifyou like, and um he's like, and
I'm like, okay, cool.
So which cases should I use?
And he's like, Yeah, therearen't really any cases out
there.
I'm like, what?
Like, I go into HarvardBusiness Publishing and I just
search on PLG and I come upempty.
(09:57):
He's like, Yeah, because it isempty.
Like, here's a case on Dropboxthat you can kind of repurpose.
Here's a case that I've writtenthat has that's not in there
yet.
And other than that, you're onyour own.
This is three years ago.
So I'm like, it's crazy becausewe actually know how to do
this, but no one's written thebook.
So that was the spark.
Sophie Buonassisi (10:17):
Incredible.
I'm glad it worked out.
Dave Boyce (10:18):
Yeah.
Sophie Buonassisi (10:19):
And glad we can talk about it today.
No, we're big believers here atGTM Fun that strategy alone
doesn't win deals, executiondoes.
But 85% of sellers are stuckmanaging their books in
spreadsheets because theirdata's scattered across dozens
of tools.
ZoomInfo Copilot Workspacefixes that.
It's the first workspace wheresellers actually work.
Complete buyer context, yourREM data, and AI-powered
(10:40):
insights in one place.
Execution wins deals,everything else is really just
preparation.
You can learn more atzoominfo.com forward slash
copilot.
It'll be in the show notes.
What are the most foundationalparts of it that you think if
you could tell any go-to-marketleader or founder listening,
hey, remember these three thingsor two things?
Dave Boyce (11:01):
Yeah.
Well, for the first thing, justto get your attention, I'd just
ask you like, when's the lasttime you spoke to a human to buy
gas for your car?
Sophie Buonassisi (11:10):
I try to avoid it at all costs and all
that.
Dave Boyce (11:12):
Of course.
Like, why would like why ofcourse?
And then you think, well, whatabout all the other stuff in my
life?
When is the last time you spoketo a human to buy groceries?
When's the last time you spoketo a human to buy an app on your
phone?
When's the last time you spoketo a human to buy a piece of
software that you use in yourwork life?
Like, it is less and lessfrequent over time, less and
less frequent.
(11:33):
That's just a long-term trend.
So then you think, all right,well, my competitors, or sorry,
my customers are versions of me.
So how much do they want totalk to me?
How much do they want toschedule a demo?
How much do they want to get onlike get on a phone call, a
discovery call?
Like zero.
They don't either.
(11:53):
And if they and if if they canfigure out a way to kind of
trial their way into successwith my competitor's product, or
just get their hands on mycompetitor's product where I'm
putting up walls, they're gonnago with my competitor.
That's just the way it works.
So then so if if we can kind oflike break out of this notion
(12:16):
that it matters at all how we'reused to doing it, because it
doesn't.
The market cares zero.zero howwe're used to doing it.
The market just wants what whatit wants and it's gonna get it
wherever it can get it.
If it's from me, great.
If it's from my competitors,great.
Market doesn't care.
So here we are kind of holdingon to the old ways.
Here's our competitor saying,Oh, you want it that way?
Cool, here, try it.
(12:37):
Then you just got to thinkabout what are the first
principles that would help youbuild in that way.
And now I'm asking, finallyanswering your question.
These are three firstprinciples in the book, and then
we go into all the details forhow you engineer this into your
product.
But the first is empathy, andthat's and unironically,
(12:57):
empathy, like literally, I wantto understand the end user of my
product.
I want to understand how shedefines progress in her life,
how she what her jobs to be doneare, and where and when it
would make sense for her to hiremy product or service to help
her make progress.
I really want to understandthat what are the words she
uses, what are them, how doesshe measure it, how does she
experience it.
Okay, cool.
(13:18):
So now once I understand that,I want to build something that's
going to kind of meet her whereshe is.
I'm not going to require her tomeet me where I am, I'm going
to meet her where she is.
And the next principle, again,unironically, is generosity.
Oh, that's the problem?
That's what you're trying.
Here, try this.
Cool, what do I owe you?
Nothing.
(13:38):
When do I pay you?
Don't worry about it.
Never.
Like, let's just see if itworks.
Like, and then once she'srolling and she's and she's
developing habit around that,eventually I will, of course.
She'll get to a point where shewants to do a longer than a
40-minute Zoom call or scorestore more than 90 days of Slag
messages or or two terabytes ofstorage or whatever.
(14:00):
She'll get to a point, shewants one access to templates or
you know, whatever.
Great.
And then she'll be happy topay.
So one is empathy, two isgenerosity.
And then the third thing is ifwe're building our products this
way, we need metrics.
Because we don't have a humanin the room here observing how
she's experiencing the product.
We don't have a human in theroom kind of coaching her to do
(14:22):
the next thing, training her,um, teaching her.
So we're relying on the productto do that work, and the only
way we know if it's working ornot is by metrics.
Like we will see the clicktracks, we will see the success
paths, we will see the deadends, we'll see where where
she's not discovering somethingwe thought she should have
discovered, and we'll see thatall in the metrics and in the
(14:43):
analytics, and then that's howwe will fine-tune our product
over time.
So empathy, generosity,metrics.
There's a whole bunch ofscience behind how you do that
and cohorted-based kind ofmeasurement, and but if those
are the three first principlesI've got to get into place if I
want to meet today's buyerswhere they live.
Sophie Buonassisi (15:00):
Mm-hmm.
Yeah, I love the simplicity andit it sounds simple the way that
you framed it, however, it issuper hard in practice.
Dave Boyce (15:07):
Super hard, yeah.
Sophie Buonassisi (15:09):
Yeah.
And how does AI come into playwith the book?
Because if we map the buyerjourney from end to end, AI is
touching more and more of it.
What is AI's impact on go tomarket right now?
Like, where is it most visiblethat you're seeing and that
you've written into the book?
Dave Boyce (15:25):
If I think about that customer journey, and in
you know, PLG terms, we're gonnatalk about um we're gonna talk
about awareness, we're gonnatalk about acquisition,
activation, first impact, thenwe're gonna talk about habit,
then monetization, thenretention, then engagement, then
(15:49):
retention, then expansion.
So these are kind of likestages in the customer journey.
Okay, cool.
So we've mapped that all out.
And then you think, all right,well, where could AI help?
It can help almost anywhere.
Like, you know, I'm let's sayI'm active, I've created my
account and now I want toactivate.
AI can can be almost like aguide to help me do the things.
It can make guesses for me.
(16:10):
I don't know if you've everused any kind of vibe coding
platforms or like an AIautomatic website generator.
You tell it a few prompts andit just says, here, let me
guess.
Here's some pictures, here's alogo, here's a um, here's a uh
login kind of um I want to saydialogue.
Um it just pulls it out of itslibrary, pulls it out of its
(16:31):
memory, does pattern recognitionand kind of fills in the gaps.
Cool, that's a way fasteractivation process than if I had
had to kind of read somethingand go do it myself.
You can think about it foronboarding.
You think about it for um forfeature uh discovery, like hey,
I'm I'm chugging along in Canva,and Canva identifies that I may
(16:54):
be able to benefit fromtemplates, or I may be able to
benefit from a backgrounderaser, or I might be able to
benefit from something that Ihaven't discovered yet.
Boom, it can suggest it for me.
Um and then you can think aboutlong-term engagement, like it
just helps me, AI just helps meum accomplish what I want to
accomplish.
Uh helps me know when I shouldupgrade to the next tier, it
(17:16):
helps me know when I shouldexpand this to team usage, etc.
etc.
The other thing that AI, sothat's just basic.
That's basic, Sophie, justbasic kind of automating the
predictable of all the wayacross the customer journey.
The other thing that AI can dothough is unlock product
categories that were previouslynot automatable.
So now I have a very complexkind of, you know, think about
(17:38):
the most complex software thereis, you know, like uh process
manufacturing ERP.
Like, oh my gosh, like supercomplex.
Um, lots of configuration, lotsof connections to uh machines,
etc.
Well, AI can kind ofdecomplexify that for me too,
like literally tutorial kind ofhelp me connect this machine,
(17:59):
help me configure this uhdashboard, help me interpret
these metrics.
And AI can literally just helpme self-serve my way into
success where I may have neededtwo or three humans and two or
three months to do it in thepast, and now a previously too
complex to be self-serviceproduct becomes self-service.
Sophie Buonassisi (18:20):
Which is incredible.
And we're seeing it happen moreand more.
I mean, you go to these Yarmywebsites, even you can see these
product led growth, these PLGflows in place.
Dave Boyce (18:30):
I mean, even S I I was just in Saudi Arabia working
with a manufacturing firm,they're doing that same thing.
Even SAP, you know, largelythought of as the most kind of
most monolithic and complicatedsoftware, they've got agents all
throughout their go-to-market,and they've got like almost a
PLG kind of flow where you firststart in a sandbox that has
their data, then you go to anoffline sandbox where you've
(18:54):
uploaded your uploaded your owndata, then you go to like a
cloud environment where you canbe more real-time, and then you
go to um like the fully kind ofpermissioned you know, SAP
instance.
But that's like a self-serviceonboarding that you would never
have thought of five years agofor SAP.
Sophie Buonassisi (19:14):
Yeah, exactly.
So what what tasks should AIimmediately own, and which ones
should humans protect?
Just it sounds like we'removing further and further along
the spectrum, which we know weare, but how should people be
thinking about that divide?
Dave Boyce (19:29):
I love it.
There's a um there's gonna be aweird way to start this answer,
Sophie, but there's a a Polishfantasy fiction author named
Joanna.
She says, I want AI to do mydishes and laundry so that I can
do art and writing.
I don't want AI to do art andwriting so that I have to do
dishes and laundry.
I really think that is how weshould think about it.
(19:51):
We want AI to automate thepredictable so that we we can
humanize the exceptional.
So anything that's Predictable.
Fill out a form, fill out afield, process an order, build a
do research on an account, youknow, respond to a trigger with
(20:12):
uh with some sort of a you knowjust a mechanical kind of
acknowledgement.
All of that should beautomated.
AI should be doing all of thatfor us.
And then the human stuff, ifyou're thinking in terms of go
to market, the the human, it'swhere we want to show up as a
real human, like the emotionalstuff, like many times on the
(20:33):
other side is a is a real human.
This is me and you talking,like if I'm the if I'm let's say
I'm with the selling company orwith the buying company, you're
trying to accomplish something.
Like you're not just out herefor fun and you've gotten a
certain way amount of the way onyour own, and now you're
trying, now you're trying tohave the courage to put this
(20:54):
forward as the standard platformwithin your company.
Okay, courage is something thatwe can work on together.
I can help you connect withpeople that have done the same
thing.
I can use my judgment or mypattern recognition and and then
and also my trust building withyou and and help you get to the
point where you're managingyour stakeholders in a way that
(21:15):
makes sense for kind of whatthey're gonna need in order to
approve that as a go forwardthing.
That's a very human experience.
It would be very hard for youto trust a robot who was kind of
trying to coach you on stufflike that.
But when we get to the pointwhere all the automated stuff,
all the predictable stuff isautomated, then that means we
have to show up as likesuperhumans.
(21:37):
Like we gotta show up likereally, really human, like
aesthetic and intuitive andhelpful, and in a how can I help
you mode, not a what can I sellyou mode.
And that's that's the versionof the future that I want to
believe in.
Well, the first thing you need,um, you need a theory of the
(21:58):
case, Sophie.
Like, um, and I think that allstarts with the data model.
Like we described the PLG datamodel from you know, awareness
acquisition, activation, firstimpact habit, you know, blah,
blah, blah.
I need a theory of the case.
I need to know what that lookslike.
I need to and I I need to havekind of mapped it out into what
I would call a data model.
Winning by design uses thebowtie data model.
(22:20):
That's the PLG version of it.
The sales-led version of itwould be um awareness,
education, selection, commit,and that's like the traditional
sales and marketing funnel, kindof narrowing as you go.
Awareness, education,selection, commit.
Commit is the narrowest part ofthe funnel.
I have it turned on its side,but it's still a funnel.
(22:41):
And then that's we that's likethe knot of the bow tie, and
then I'm gonna start opening outfrom there.
Onboarding, retention, andexpansion is gonna hopefully
kind of make my long-termlifetime value with that
customer actually expand overtime as we deliver impact.
Okay, now if I can define eachof those um stages and I can
(23:02):
define what needs to happenwithin each of those stages, you
can imagine that that is like aum almost like a manufacturing
process, right?
Something happens during thisstage, something happens during
this stage, there's a handoff,something happens during this
stage.
There's success criteria,there's activities in stage,
there's success criteria to exitthe stage.
And then once I have it definedlike that, that's kind of like
(23:23):
bare minimum, and it's not alot.
Like that's bare minimum for meto be able to start running
experiments.
Then I can start A-B testinghuman versus robot on this task.
Did it help me improve myconversion rate?
Oh, it didn't?
Okay, then that's probably notthe right place to be using a
robot.
How about or maybe I tune it alittle more and see if I can get
it there, tune it a littlemore, eventually I get it there.
Oh, cool, cool.
(23:43):
The robot can help me there.
It helps make my reps moreefficient, it helps make my
conversion rates better, but Igotta be able to measure it,
which means I need a data model.
And that data model needs tonot be a political data model.
I'm sure you've been in boardmeetings, I've been in uh board
meetings where it feels like I'mjust on the receiving end of
like a commercial and theoperators are just trying to
(24:04):
kind of convince me that nothingstinks in this business.
That ain't that's not what weneed.
We need a very clear light ofday, consistent data model so
that we can run experiments,because we're just gonna run
experiments and we're gonna getAI deployed everywhere that it
works, and if we deploy it andit doesn't work, no harm, no
foul, we'll kick it out andwe'll try it somewhere else.
So I think data model first,kind of theory of the case.
(24:26):
What are we trying toaccomplish?
And then we can start choosingwhere we're gonna deploy AI, and
we'll treat it as an experimentuntil it's proven that it's a
permanent thing.
In lean manufacturing, this islike one of my favorite things
about uh sayings in leanmanufacturing that you never
hear.
You always hear aboutcontinuous improvement or the
and-on cord or uh just in time,but here's here's a very cool
(24:47):
one.
Don't bolt down what you can'ttape down, and don't tape down
what you can't hold down.
So basically, I don't need togo bolting the AI in place when
I still don't know if it's gonnawork.
Like, let me just hold it inplace.
Let me test it.
Oh, that looks like it's gonnawork.
Now let me tape it and let mestep away for a second.
(25:09):
Oh, it still looks like it'sholding.
Okay, now we're convinced AIcan do that job consistently,
reliably, cool.
We're gonna bolt it down andmove on to the next experiment.
Sophie Buonassisi (25:18):
What does the bow tie model actually expose
that traditional funnels hidewhen winning by design is
employing it?
Dave Boyce (25:24):
That's where the renewals happen, that's where
the expansion happens, andthat's where the growth loops
are initiated that can pull newcustomers into the front of the
funnel based on referrals fromexisting customers.
All the compounding happens onthe right hand side of the bow
tie.
Now we spent, if we if you grewup when I did or anytime, you
know, if you built a companyanytime before the last five
(25:47):
years, you might have spent aton of time in QBRs and planning
sessions and board meetingstalking about calling and
meeting your bookings forecast.
That's basically what WallStreet tracks.
That's basically what you knowmost sales-led organizations
track.
It's where we put our mostexpensive people, it's where we
(26:09):
put all of our executiveattention, it's where we put all
of our focus in those meetings,is calling and hitting a
bookings forecast.
But bookings is the knot of thebow tie.
It's the beginning of thejourney.
Everything after that is thecustomer's experience, and it if
that customer experience isgood, then renewals and
expansion will happen, whichmeans now I have the machine
working for me instead of meworking the other way around.
(26:31):
So why wouldn't I also bespending time and attention
there?
Why do I put all of myexpensive time and resources and
people and attention onbringing new customers in and I
short sheet the right-hand sideof the bow tie?
It just doesn't make sense froma systems or math perspective.
And uh and the bow tie kind ofjust like brings that to life
because once you start runningcohorted math through that
(26:53):
system, you start seeing like,oh my gosh, this short term,
yes, short term, I will get alot of benefit from sales, but
long term compounding growth isall driven on the right side of
the bow tie.
Sophie Buonassisi (27:07):
Interesting with those cohorted systems are
real.
So, what would be if we take uhan MVP, for example, for
inspiration on the product side,what's like a minimum viable
bow tie for a company to need tobe able to run more of a hybrid
AI and uh PLG or go-to-marketmotion?
Dave Boyce (27:25):
Yeah.
The hardest thing is connectingthe right and left side of the
bow tie.
So all almost all of us havethe left side of the bow tie
instrument in some way, shape,or form.
We have stage one, stage two,stage three opportunities.
We've got it all in HubSpot orSalesforce.
We kind of we we kind of knowhow to build our kind of
bookings forecast and we manageit, and we have MQLs and SQLs,
like we have that built.
Sophie Buonassisi (27:45):
Yeah.
Dave Boyce (27:46):
Where does the stuff on the right hand side live?
Sometimes it lives in somethinglike a gain site, sometimes
we've written it back to ourCRM, and sometimes it's neither
place, and you have to go get itout of finance.
It's literally tracked based onbillings because we don't have
it in CRM and we're not trackingit in a CS platform, and we
literally have to say, well,when did we, you know, did we
(28:08):
send that customer a bill or notbased on whether they canceled
or not?
So minimum viable product wouldbe getting the left hand and
the right hand side stitchedtogether all the way through the
journey that I can see in oneplace.
Once I have that, and if I'vedone it according to the winning
by design uh bow tie dataschema, then I can benchmark it.
(28:28):
We've got 300 companiesbenchmarked, and you can cohort
that based on companies that aresimilar to you or on similar
motions to you, and then you canstart benchmarking.
But even if you can'tbenchmark, at least you can
compare yourself to last periodon the period before.
Now you've got like a baselineand you can start seeing if if
you're actually makingimprovements.
But before you have that, Idon't think you can improve a
(28:50):
human system.
And I also don't in a reliable,consistent, kind of ongoing way
and like a continuousimprovement in lean
manufacturing way, and youcertainly can't improve an AI
system or human AI hybrid systembecause you just don't have the
instrumentation to tell umwhat's working and what's not.
Sophie Buonassisi (29:07):
So when teams are wanting to create this bow
tie framework and they need toeither take their data from
GainSite or some other system,where are they creating the bow
tie?
Where's this connectivitybetween the left side and the
right side of covering?
Dave Boyce (29:22):
So ideally you would write everything back to CRM.
That that's e way easier saidthan done.
CRM wasn't necessarily builtfor things like growth loops.
Um very tough to do that.
Certainly not built for thingslike um you know activation of
an account pre um pre-payment.
(29:43):
Um that's a PLG thing, likevery difficult.
But let's say you could get itinto CRM.
That would be my choice A.
Um, but what we very often seeis we'll get it into like
Snowflake.
Um or we'll get it intoSnowflake and then a
visualization layer like a umlike a Domo or uh uh you know
(30:05):
some sort of or Power BI or somekind of visualization layer.
There are also some uh productsout there that are doing this
commercially um that arepartners of Winning by Design
that that will take that willvisualize the bowtie for you,
like a UNA or a Vasco.
So like SaaStrack is a partnerof Winning by Design, adheres to
the bowtie data schema, buildsthe custom objects, it's a it's
(30:26):
a managed package, builds thecustom objects inside of
Salesforce for you, and now youcan just kind of manage it and
use Salesforce Reporting.
Amazing.
That's a SaaSTrek.
Vasco and Una pull it all out,like if you have it in disparate
systems and they give youvisualization and a management
framework.
Or you can do it like I said,roll your own and put it in your
own BI tool.
But you do want that wiring tobe um you want it to be wired,
(30:51):
not just kind of like CSV kindof one-time pulls.
Because you track it thismonth.
The continuous exactly.
Yeah.
Sophie Buonassisi (31:00):
Right, right.
Super, super.
Who owns that process?
Dave Boyce (31:04):
Jeez, you are so mean.
You're asking all the hardestquestions.
Ums, it's not very consistent.
Ideally, you would have afunction called RevOps that
didn't just work for sales.
And ideally, that RevOpsfunction would build those
(31:25):
systems and would be the kind ofimpartial, objective arbiter of
truth.
In many companies, that's notwhat RevOps is.
In many companies, the RevOpsdoes whatever it takes to make
the CRO look good.
And they help them prepare forboard meetings, and they help
them kind of scrub and whitewashnumbers, and they help them
kind of um maybe maybe uh inprivate they're look looking for
(31:50):
truth, but in public they'rewhitewashing.
That ain't gonna work.
That's just not gonna work.
So ideally we just we just stepRevOps up into a kind of
impartial arbiter of truth, andif not, then it can also be
FPNA.
But the problem with FPNA isthey don't actually understand
the go-to-market well enough andthe systems that run
go-to-market well enough.
So my ideal, Sophie, would bethat RevOps steps into this role
(32:14):
going forward and is thebackbone of our modern
go-to-market.
Sophie Buonassisi (32:19):
I love it.
And for other operators, notjust RevOps, I have heard you
say that.
They need to become more chieffigure it out officers, which I
think is a really fun term.
What does that role look likein the day-to-day?
What does that entail?
Dave Boyce (32:35):
Full attribution.
That's Ryan Sanders fromMercado.
He gave me that term, and I'vebeen I've been shamelessly
reusing it.
Chief figure it out officer,it's super easy to remember.
But it's all the stuff that youand I have been talking about,
like where would I deploy AI?
I don't know.
I can't pull that out of my bagof tricks.
I can't go, you know, rewindthe clock 15 years to when I
(32:58):
used to be an AE and tell youhow I used to use AI.
I didn't use AI, which means wegot to figure this out
together.
I certainly didn't use itsystematically.
I can't tell my RevOps personwhat we did 15 years ago to
instrument or go to market sothat we could run A-B
experiments with human robothybrid systems because we didn't
have that.
(33:18):
So we got to figure all of thisout as we go, which means if
I'm if I'm a veteran, which Iam, um if I'm a veteran head of
revenue, I gotta get out of theidea that I'm just gonna teach
people how to how to do it theway I did it.
And I gotta get into themindset of no, no, no, we're
(33:38):
gonna figure this out together.
So I'm gonna grab smart peopleon my left and right, we're
gonna go in, we're gonnasystematically architect
something that I can then run byinstrumentation versus running
in an Amelia Earhart way, like,ooh, it looks cloudy over there,
I better steer left.
Sophie Buonassisi (33:54):
So you've obviously spent a lot of time
writing your own book.
Are there other booksthroughout your career that have
made a particular impact onyou?
Dave Boyce (34:00):
Oh, good question.
Yeah, I've I my my author herois Clayton Christensen.
Um he's amazing.
I've read, I think everythinghe's written.
I think he is he's a reallygood scholar.
Uh may you rest in peace.
Um I went to both of hisfunerals.
He was a good friend too, but Iknew him as a scholar and a
friend.
And um, so what I alwaysrecommend if you haven't read
(34:23):
anything of his is CompetingAgainst Luck.
It's not his most famous book,but it's a really good book for
this moment.
Um and that's where he reallyunveils the jobs to be done
theory.
Uh I super like um RogerMartin's book called Playing to
Win.
Um very, very good strategyframework.
(34:43):
Um I could go on for for daysabout books that have made an
impact, but that's where usuallywhere I start.
Sophie Buonassisi (34:52):
Very cool.
Great recommendations.
And what about yourself?
Where can people follow alongyour journey?
Obviously, we'll have a linkfor the book, but for yourself
at all, are you on LinkedIn X?
What's the best place to followyou?
Dave Boyce (35:02):
I'm not active on X.
I would love for obviously youcan find me here, and we're
gonna put that in the um shownotes.
I'm on LinkedIn, pretty activeon LinkedIn.
Uh I have a Substack uh whichis just Dave Boyce.
Um and uh and if you join theGrowth Institute with Winning by
Design, you'll see me superactive there.
I run quarterly case studies,MBA style, kind of executive
(35:26):
education case studies.
You'll see me on stage at thesummits, and we will uh and
we'll go change the worldtogether.
Sophie Buonassisi (35:34):
Amazing.
Amazing.
Dave, this has been phenomenal.
Thank you for the time.
Thank you for the book on mybookshelf.
Dave Boyce (35:40):
Yay!
Sophie Buonassisi (35:40):
And uh yeah, really, really appreciate it.
Dave Boyce (35:43):
Thanks so much, Sophie.
Amazing.
Let's go do it.
Sophie Buonassisi (35:46):
Let's do it.
Popular Podcasts
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.
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.
Two Guys, Five Rings: Matt, Bowen & The Olympics
Two Guys (Bowen Yang and Matt Rogers). Five Rings (you know, from the Olympics logo). One essential podcast for the 2026 Milan-Cortina Winter Olympics. Bowen Yang (SNL, Wicked) and Matt Rogers (Palm Royale, No Good Deed) of Las Culturistas are back for a second season of Two Guys, Five Rings, a collaboration with NBC Sports and iHeartRadio. In this 15-episode event, Bowen and Matt discuss the top storylines, obsess over Italian culture, and find out what really goes on in the Olympic Village.