Subscribe
Share
Share
Embed
"By definition, most marketing automation platforms are creating silos, not breaking them down. They're in the business of creating copies of data." — Peter
Peter Oleson spent years inside the machine. Salesforce Marketing Cloud architect at Merkle, then Iterable, running live demos of the exact journey builders he now argues are the wrong model. Today, he's making the case that marketing automation platforms should stop trying to be the brain and start being a better pipe.
That is a specific, testable claim, and Jacqueline holds it to the fire – oh, and this is just Part 1. Part 2 next week.
The central bet is worth stating plainly: Marketing Automation Platforms (MAPs) should stop trying to understand the customer and instead become the system that reliably executes a decision made elsewhere. If Peter is right, an entire category of enterprise software loses its core value proposition. If he is wrong, he is selling a solution that requires a seven-figure engineering headcount that never appears on any vendor invoice.
This episode covers accountability in the age of AI agents, the real cost structures of composable versus bundled platforms, why "breaking down data silos" has always been more of a sales pitch than an architectural reality, and which move by a modern MAP would fundamentally reshape the competitive landscape.
Timestamps
03:12 — The Sendomatic 5000: marketing automation's humble beginnings
11:40 — Vote with your feet: why talk is cheap in martech
13:00 — Before AI agents, humans were already getting it wrong
15:19 — You can't abdicate accountability to AI agents
21:45 — Composable martech finally unites lifecycle and performance teams
25:14 — Data portability isn't a tax. It's an investment with dividends
37:50 — Who becomes the first modern "dumb pipe"?
41:40 — Unify the data, personalize the moment: why every martech pitch sounds the same
49:04 — When the data silo is your only lifeline
Sponsor
Brought to you by Hightouch, the leading composable CDP and decisioning platform trusted by brands like Domino's, Chime, and Aritzia. 90% of customers have a real use case live within their first week, delivering world-class personalization at scale. Learn more at hightouch.com/msom.
Connect & Subscribe
Creators and Guests
What is Making Sense of Martech?
Unfiltered takes on the biggest shifts in marketing technology. We spotlight what matters, who's leading (or lagging), and what's next. In Martech, clarity is power — and we're here to deliver it.
Welcome to the making sense of MarTech, where the rabbit hole goes deeper than the headlines. This
is a show that pressure tests ideas, not just platforms. Peter Oleson spent years using
Salesforce Marketing Cloud in an iterable. Teaching customers how to build the exact journey.
Builders, he now says, are architecturally broken. Today we find out if that's a
conviction or a business model. By the end of this conversation, we'll know if Peter is a plumber
telling the truth about the pipes or a guy selling wrenches. A little bit about him first.
Peter is a solutions engineer at Hightouch, the composable CDP and authentic marketing platform
who's betting on the future. The idea that the map category he came from is running out of reasons
to exist. He didn't arrive at this argument from the outside. He built cross-Channel journeys at
Merkle as a marketing cloud architect, then moved to iterable, where he ran live product tours,
specifically teaching customers how to build the automated journey builders he now calls the wrong
model. And that's not a career pivot by any stretch of the imagination. That's complete
reversal. And from an insider. Welcome, Peter. Thank you for being here. Thank you so much for having
me. I'm a big fan of this podcast. So this is a this is an energy your energy. And I'm a long time
viewer. I am the YouTube viewer. Yes, I'm grateful for that. And also we need you to go ahead
and out of the way. Put a disclaimer, at least for my end, I have all editorial
independence, and so this has absolutely nothing to do with the fact that Peter works at High
Touch. And I think, Peter, you probably have your own equivalent. Oh yeah, I have my own martech
hatch act to say here, which is that everything that we talk about, all the opinions that I have,
these are my opinions. They're not the opinions of Hightouch or any former employer that I've had
for that matter. I arrived at these all on my own. So these are mine? Yes. and I have receipts.
If people really get upset, this conversation Peter and I've been having for at least 4 or 5
months. And so please note, this is just truly a philosophical
conversation, and I'm so grateful that you even came up with it, because it has rocked my world
for the better part of this year. I'm excited to have the conversation. Me too. At least in a
formal capacity. That's right. Let's start off with some rapid fire. What was your
first martech tool? So technically. And I had to reach out to a former colleague. This
was Dreamweaver and an in-house built tool called the cinematic 5000. Um, I
think it was like a Saturday Saturday Night Live bass reference that's dating me. Um,
but, uh, it was really plain. You could basically do some minor audience
building. It didn't even have scheduling, so I was like waking up at seven in the morning to hit
send so that on the East Coast they would get it. Um, but in terms of enterprise grade, that was
Marketo. And at the time it totally blew my mind. I was like, oh, you mean that you can capture events
directly on site and do follow ups? This is amazing. Sign me up. Understood. Also, I
feel like it's it's no longer cool to say, but I love Dreamweaver. Oh, yeah. It's great. Like
anyone who's been in the game long enough knows that. Like, if you want properly rendered emails,
it's got to be HTML code. Oh, yeah. And also just Dreamweaver was the first of its kind to make it
easy to do both visual and code at the same time and see what you're doing in real time. Change the
code over here. See it over there. It was great. Exactly. All right. Real talk. Garbage
disposal. Water heater or dumb pipe? Which kitchen appliance are you? Do I have to choose one
of the three, or can I choose my own? You can choose your own. Uh, if I'm a kitchen appliance, I'm
probably something more like, um, like a French press, which is not like a sophisticated, like,
mechanical appliance. Uh, but the reason I say that is that, like the users of a French press,
like, pretty opinionated. I'm pretty hands on. Um, and I'm also a firm believer in,
like, better inputs and thoughtful processes produce a much better output. So, like, good beans,
good grinder and ten minutes, as opposed to you flip a switch. Um, it's
going to produce a better output, better cup of coffee in this case. I like that. And also you work
well under pressure. You've held three distinct seats in this industry ESP map vendor
agency selling the platforms and now a warehouse. Native company arguing maps are
broken. Which version of Peter in your whole career was most wrong? So I
choose to say that like, I learned new information and changed my mind rather than being wrong. Um,
but if I have to choose, it's probably either agency. Peter or. I spent a very,
very brief time, um, as an account executive at a map. Um, and it was just not the
right role for me. But agencies are great. They make a ton of sense for specific types
of companies. However, like, in my experience, they're kind of always after
the almighty billable hour and increasing those. Right? Um, so there can be at times a little bit of
a conflict of interest there, right? Because if I'm in the business of selling you more hours of my
time than potentially, I can get into this dangerous territory of being self-serving rather
than serving the customer. Right? Yep. It's as if that's Deloitte and Accenture's home model.
But. Right. And I'm sure that there's there's some people who are going to listen to this who employ
agencies for a specific part of either their marketing ops or their strategy. And I'm sure that
there's a subset of those people who are like, oh, yeah, that one campaign that's super complicated. I
actually don't know how the sausage is made. Like, I don't know what it actually takes to execute
this. Um, and I think that that's pretty normal. And so what I would say there's like, I've always been
invested in, if you're going to go the agency route, if you're going to bring on consulting, make
sure that, yes, they fish for you, but they also teach you how to fish, because that's the
critical part, because that relationship will end at some point, and you need to know what to do
from that point forward. Yes, nothing is worse than all of your company's institutional knowledge not
being in-house, right? And that's not just a blueprint like that can't just be a blueprint. You
have to get hands on. You have to understand what's going on there. Exactly. Okay, so if a brand
came to you tomorrow and said they can't afford to have 3 to 5 dedicated engineers, would you tell
them to buy an ESP marketing automation platform? Sep. Whichever name we're going off of the
alphabet soup. This industry did not make it easy on anyone. Um, so. All right. Would I tell
someone to buy a marketing automation platform if they had? Let's call it limited resource on the
data and engineering side. Um, as someone in solutions like my skin kind of crawls answering
this question without, like, asking five follow up questions, but I guess I'll do my best here. I
think you can get away with fewer data engineers than you think if they're
fully dedicated to the marketing team, and I say the marketing team and not just the platform,
because oftentimes there's other platforms that are involved here, right? So I think you could
probably get away with fewer than five. But what I would say in addition to that
is the most important thing that I've seen is that regardless of the number of data engineers
that you have, right, is that you do have to have one or maybe two people that are the ultimate
decision makers, because this isn't a democracy. That's something that actually a former boss of mine
used to say a lot is like, this isn't a democracy. We need to make decisions. Democracies
take too long to make decisions. So you do need to have some 1 or 2 people
that give a go, no go at the end of the day, and they have to be accountable for that decision. Um,
another thing that I would say is I don't actually think in the future that data
engineers should be in charge of building the pipes between your data platform and then the
tools that your marketing team uses. So like could you could like snap accordingly. Like.
Yes. Yeah. I just like that's not a good use of time. And I think it's also
important to note that the more time that your data engineers are spending in the data
platform, that's actually a net benefit to marketing, because then you get to benefit
from all the amazing data engineering work that they do, as opposed to, hey, can we get this one
additional attribute over here in Salesforce Marketing Cloud? Like that's just when you
consider this an even dumber pipe. Uh, what do you mean, there?
I mean, having a data engineer set up these pipelines with likely minimal
observability, minimal understanding of what they're building. They're just doing point A to
point B, as they were told. Definitely a dumber pipe than the the execution layer inside of the
map that we've talked about. That makes sense. Okay. Are you still friendly with your former employers?
Of course. Yeah. Um, this is a small industry. Like, despite being, you know,
10 to $25 billion, uh, market, depending on how you look at it. Like,
it's a small group of people. And so you run across the same people a lot. Um, and so, yeah, I'm
friends with folks from from my agency days, from my marketing automation platform days we
interact in public on things like LinkedIn. We interact in private. Like, I still have text
threads with former colleagues of mine in the space, um, a few of whom reached out after I
published the essay on LinkedIn. Um, oh, I met. Yeah. It's I mean, it's all love. Like, it
all comes from a place of I want to see this industry as a whole succeed. And in order for us
to do that, we have to really have a close ear and understanding of
what the customers who use these products want and need. And so that's really what this was about.
Um, but it's also worth noting, like at the end of the day, I'm such a product nerd that
I vote with my feet on that front. So, like, if I have an opportunity to go and chase what I think
is going to be the next thing. I'm going to go vote with my feet and go do that. Understood. Okay,
so when an agent is making the send decision and it's wrong, who gets angry? Slack messages
the marketing automation platform, the warehouse or the agents vendor? I mean, short answer is
probably wherever the agent lives. Like, maybe that's in the warehouse, maybe that's in the
marketing automation platform, maybe that's somewhere else. Um, so it's probably wherever the
agent lives. But I think we need to be clear that before agents made
these decisions, humans did. Right. Yes. Um, and to a certain extent, it was a lot worse when
humans made the error. Because oftentimes, if we're making decisions inside of a marketing automation
platform, this is like a multi week sprint situation where we're trying to uncover
every possible thing that could go wrong And mistakes still occurred, right? Um. Oh, yeah.
So, um, I also think that there's another thing to think about
here, which is did the agent have enough context and understanding to
make the right decision? Right. Um, and this can be does it have just enough raw
material to work with? Um, so if it's inside of the marketing automation platform, you can probably
tell my thought processes, no, it does not have enough context to work off of. That is a subset of
the overall context that it could and should be using. Um, on top of that. Does it have an actual
like the term that's kind of an eye roll. It's like semantic understanding of the data, basically.
Like can it read the data based upon what you tell it that data means? Or is it having to infer
all of that? And I guess what I would say as by way of advice is like, if you're employing
agents that don't have all the proper context and the understanding, then you probably
need to take a little bit of a look inward and like maybe you're sending yourself the nasty
slack. Right. Yeah. I think a lot of it is kind of like self-driving cars. Is it
the driver? Is it the car? Is it the car company? Is it you? There's just too
many nuanced question marks that we haven't quite figured out. And I think it's the same concept.
Totally. What laws or what regulations can we internally impose to
maybe avoid these types of things and circumvent. Right. And that's why for a long time like you,
you're still going to have to require that human behind the wheel of the car. It's no different in
marketing. And so if you're not reviewing the outputs or the decisions of
whatever agent you're employing. Then I would say, like the
nasty slack message absolutely has to go to you. Like you don't get to abdicate responsibility
just because agents have gotten better in the last two years. Yeah I agree. All right. Last but
not least, what is your hottest take? You're gonna hate this one. Oh,
um, if we're talking traditional marketing automation platforms, I don't think
anyone beats the original exact target Salesforce Marketing Cloud. I don't hate it. I just
dislike the fact that it's true. And here's my reasoning. Like, they're probably going to be
people that's like, this is the guy who's talking about the dumb pipe and he's talking about exact
target. Can't get data in fast. And, uh, you know, the contact model still exists. Like, yes, all of
that's true. But if we're if we're having to choose on traditional maps, um, I don't think it
gets beat. Um, I never worked in Salesforce Marketing Cloud next. I'm not a big fan of like,
all right, we have to have CRM or data 360. Um, but the reason why I think that
Marketing cloud, the original exact target version is the best is because while it takes
a really technical person to make it run the way that it needs to, if you have that
person, it's infinitely flexible. You can do things that modern platforms
can't do. Um, things like, oh, every modern platform has some kind of eventing in it, right?
But if you want to say, all right, I want to take the most recent three events that were attributed
to Peter. So like, maybe that's recipes. I'm a huge cook. So like, I want to take the last three
recipes that Peter looked at on this site. And I want to surface that in a blast email. Like if
that's in your data extensions, you can do that in Marketing Cloud, whereas a lot of the modern
platforms are going to be like, oh, you want to do that? Huh? Okay. Well, our recommendation would be to
aggregate that outside of our platform and then just send it into us. So it shifts all of the work
to say, all right, you have to do this. You still kind of have to do it inside of Marketing Cloud
two, but you can do it with AMP script. If you have someone who knows what it is that they're trying
to achieve. So that's why I say that, no, I agree, I still stand by in marketing. Cloud is the most
powerful tool, but the hardest at the same time. You really have to
have an architect overseeing it 24 over seven. Yep, a lot of things can break there. It's really
expensive. But if like cost is no object, it's a pretty good tool.
Um, I'd also probably say, I don't think this is a hot take really. But I'd go one step further,
and I'd say that a lot of the next generation of marketers are going to be worse off, having not
been forced to learn a platform like a marketing cloud, like a pardot, etc.. I agree, I
think, um, in terms of the technicality, you can still be a great marketer, but in terms of
problem solver and solution of okay, how can I do this in a different way? There's just a
lack of flexibility. If you haven't had to build something that it was not prepared to do.
Right. And then also like we talked about in your last question about about agents is like you have
to be able to evaluate the output, understand what it's doing and know where it has gotten things
wrong. And I think if you've been in one of these platforms that was kind of your full time job for
a while. And so the diagnosis and identifying, uh oh, this is maybe where something
went wrong, but let's flag a couple more places and then do the deep dive. Yep. Totally. Um,
okay, let's set the stage. So your central bet has been. And it's worth stating very,
very plainly before we test it. Map should stop trying to understand the customer and become the
system that reliably executes a decision made elsewhere. Bigger pipes, not
smarter platforms. And the stakes are really present and real. Because if you're right, an
entire category of enterprise software loses its core value proposition, and that's been a mainstay
for a long time. And if you're wrong, you're selling a solution that requires seven figure
engineering teams that most companies don't have. And so you argue that maps charge
customers to store the data that already resides in the warehouse. And it's a data tax. But the
counter charge is that the composable alternative sometimes costs, you know, anywhere from half 1
million to 1 million a year in data engineering headcount that never appears on any vendor
invoice. Is that just a different tax with better PR?
I don't think it's a complete equivalent here, and I'll try to explain why. So
like the data tax is just the dumb part of it. Um, you also pay for activation and execution, right?
Like every platform out there has a CPM. Or if you're marketing cloud, you have super messages,
which I'm still not sure exactly what they are. They still don't know either. Also, like if you
want to get a laugh out of anyone that has used marketing Cloud, just ask them what a super
message is. You pay for the data storage, you pay for the activation and execution.
Um, you're also still in the case of maps, like you still have data and engineering headcount today.
Um, and so there's maybe a question of like, are we talking about going and building
in-house? Like, that's certainly not free if we want to go that route. Um, it's also not free if
you want to go. Well, in the case of an in-house build, you're probably going directly to an MTA. Um,
but maps charged you to store a sliver of the data,
right? Like, it's just a portion of the data that already resides in the warehouse, but with a more
composable approach like the date, as the data model changes, as, um,
it grows or contracts or changes in its nature. Like, I'd argue that in a composable
approach, having access to that data is worth some kind of tax, right?
The other thing that I would say is like you're able to execute off of more of the marketing
surface, right? This is something that I'm hearing a lot in conversation now is like, how do we align
our life cycle and our performance marketing teams together, which is not really a strong
suit of the traditional map right there, mostly focused on own channels, email, SMS, push.
Um, you know, in app. I guess what I would say is like nothing's free.
Even in-house platforms cost time and money to build campaigns and cost even more money to build.
And yeah, that's a sad reality. I mean, I think we're gonna see in, you know, a couple of years, all
of the people who vibe coded their way into a marketing automation platform are going to come
back. Um. Oh, yeah, it's it's going to be a mass exodus. Yeah. Right. Um, but I would
say, like, the composable approach gets you a clearer picture of your customer
with the ability to do more things with that information,
and you buy what you need, not a bunch of things that you don't, right? So, like if you're buying a
map, like you're automatically usually getting the audience builder, the campaign builder, the journey
builder. Um, other, uh, insights tools that you may or may not use, like all of that goes into
your cost. Um, but with a composable approach, like you can kind of pick and choose what parts you
need and what parts you don't. Yeah, I kind of think of it as the the old, like, cable bundles.
It's like, I actually don't want all 800 channels. I want these 20. I
only want to pay for these 20. And if you don't want to pay for the remaining amount that you
will never use, That is the difference. And that is what you need to be able to laser
focus on. And so yeah, hopefully we kind of expanded and blew out with
the um, with the streaming platforms as well. Right. Like I it's coming back. We're bringing back
bundles essentially. Yeah. We're going to consolidate. There's going to be consolidation
there. But I guess like yes, there is a tax in terms of data and
engineering headcount. I guess a major difference there is all of the work that data and
engineering does is in your data platform with your data and you own it. Like that's a that's
a benefit. And so when you actually go to switch platforms or something like that, you take that
with you. It's not stuck in a marketing automation platform. And you now have to have that as part of
a problem during your implementation and your switching. Yeah. Truthfully, I would say it's not a
tax. It's an investment with dividends. That's a good way to put it. Are you in marketing?
Maybe I don't know. Yeah, that was good. Um, okay. In your essay,
the RFP language you reference vendor must support composable architecture is a
bit anonymized and a little generic. How often does that clause actually survive a six month
implementation? When the team realizes they need to stand up and maintain the pipeline themselves?
I would say it's like a borderline zero hit rate when you're looking at a traditional map, right? Um,
and there's a lot of reasons for that. Like, and not all of them are related to the marketing
automation platform. A lot of them are related internally to the teams who are asking that
question. Um, but why I say 0% hit rate? And by the way, like it's worth
noting that like, this isn't really thought leadership. This is like, these are things that
customers come to me with and say, like, this is important to me. And so this is very
reactionary. And and what I'm seeing in this space, not necessarily where I think it's
going. It's already here, but why this is a 0% hit rate on all of
the, um, the implementation side of things. Part of the reason is because the data
foundations for the marketing automation platforms that these customers are asking the
question to don't truly support it, even when they market themselves as such, like
every platform, to my knowledge, has some concept of the user profile.
There is maybe one platform that I can think of off the top of my head that does a true
warehouse native version of this where they're not storing it. But then there's all sorts of
other problems with latency and scalability there. Um, I'm not going to name names there. It's not
particularly important. Um, you can name it. We can cut it because I just want to know. Yeah. That
makes that that is a warehouse native platform that data and engineering teams love.
And marketing teams tend not to love as much. Correct. Okay. That tracks so like
they have this profile and they're forcing you into that. Um,
and they say things like we can connect directly to your warehouse, but that's warehouse
connectivity that's not sitting on top of your warehouse. You're still transferring that stuff in.
Um, so part is on the marketing automation platform side. The other part is actually on the
organizations who are asking that question, because they often put that in an RFP because
that's been like the data posture for their organization. But two months down the
line, they're ultimately going to say, well, I have this one specific use case. I actually want to
build the audience inside of the marketing automation platform and not outside of it, just
because it's easier for some reason today, right? Maybe long term that's not the case. But the
second you say, all right, I need this, but there's this one snowflake use case over
here. That's a unicorn. You're talking out of both sides of your mouth. So I think like it's
important and it's incumbent upon the the organizations asking that question to be
really honest with themselves. Is this a preference or is this a need brought to
you by our sponsors? If there's one theme that's followed me at every stop in my martech career,
it's trying to get good data into the hands of marketers. That's why I'm so excited to tell you
about our sponsor, Hightouch, the leading composable CDP and AI decisioning platform
companies like Domino's, Chime, Aritzia and PetSmart trust Hightouch to power their data. And
here's the kicker 90% of customers have a real use case live in production within their first
week. That means you can implement a world-class CDP in months rather than the usual years long
headache. That's why top brands choose Hightouch to personalize every customer interaction at
scale. See what Hightouch can do for you at Hightouch.com. And now back to the hot seat. Very valid.
I mean, I'm always gonna say I think segments should exist in a map, but they're based on what's
in the data warehouse. Exactly. Because you want that easy accessibility. But if you're doing
omnichannel, multichannel, if you're thinking of decisioning, there's a lot more that goes into
play. And you have to really kind of take a step back and figure out, do you need to go upstream? Do
you need to stay where you're at? What works? What are the capabilities? Pros. Cons Name it. Totally.
I'd throw out there as well that, to my knowledge, there are zero marketing automation
platforms that will allow you to disable things like profiles and audience building because it
happens elsewhere, and then also receive a contractual benefit in a reduction of your total
platform cost as a result of not having those. Right. It's like number one, they don't do it.
Number two, you're not going to save money by requesting that they not do it. So I
can see if I can try. That sounds like a challenge. You're probably a better negotiator than I am. Uh,
for better and for worse, either your best friend or your worst nightmare. Right? In
your piece, you're projecting 30 to 50% of enterprise adoption of this new model in the next
2 to 5 years, starting with elite companies and filtering down yet Iterable. Braze Klaviyo
customer I know they're all still growing today. At what point does rising RR stop being evidence
you're wrong and start being evidence. You're just really, really early. Well, I definitely think I'm
just really, really early, as you well know. And probably everyone watching this knows enterprises
don't replace core marketing infrastructure overnight. Like this is a slow
burn. Um, and they renew contracts even when it doesn't serve them. Um, or if they
don't remember, it just auto renews, right? Yeah. Crazy to me. There's the whole auto renew. Make
sure that, um, you know, that's written into your contracts around, like, there needs to be a
notification on both ends, right? Um, but, uh, they renew contracts. Sometimes they auto
renew. Um, but there's been a really gradual movement of data
decisioning, orchestration upstream for a while. Right. Um, some really
large customers that I've worked with have taken the approach that like, hey, these features inside
of the marketing automation platforms. Um, and this is agnostic of platform. Uh, I don't
need these anymore. They actually hurt me more than they help me. Um, and what I've seen a lot of
is, okay, we're just going to use your trigger API endpoints, and every platform has them.
We're going to use those because that gives me more flexibility and
functionality and upstream orchestration. Then do sending in an event
to a braise or sending an event to an Iterable or a Marketing Cloud. So
the early signal I don't think is declining IRR. It's more
shorter Commitments like the one year commitments. Um, it's more
API driven execution or, you know, MCP using an AI harness.
Execution. Right. It's it's when you start to see inside of these marketing
automation platforms that they're not using the core functionalities that these platforms were
built on, things like journeys. And when you start to hear more
questioning of your customers as to what they're actually paying the ESP to own, and I
feel really deeply for my customer success compatriots out there because they're the ones
that get faced with this. All right. We've been on this platform for 2 to 3 years. Um, when we
evaluated it, we were really interested in journeys. But now, for reasons kind of outside of
the marketing team's control or Intentionally. On the data side, we've just shifted our internal
position and we actually get better results by going direct to warehouse. Yeah, that's kind of
where my head's at there. Yeah. It's a we're at a tricky kind of impasse of it's all
a philosophical choice and what each company is considering. And the same goes with like build
versus buy, which one is the cultural backbone and which one actually
aligns with the methodology. Totally. And I don't I don't think it's like it's not really a question
as to whether or not ESPs or Maps have value. It's like, where does the value come from?
Like, does it come from owning the customer model or does it come from being a totally
awesome and execution layer in the stack that like, you never have to worry about? Yeah,
exactly. I think it's going to be the latter. But yeah, at the end of the day, I mean maps,
esp, EPs, they're all just UI and UX versions of sending on an MTA and
standard email protocol. Yep. But it's also needed because not everyone is technical
enough, nor really should they have to be. And that's where a Wysiwyg and or a code,
you know, a coder to be able to HTML code your email. Like there's pros and cons to all the
things. It might be cheaper to go to a different route, but it doesn't always mean it's the right
choice for the technicality of your team. Yeah, it's completely team dependent. Um, and
on the map side, one other thing that I would call out, I think it was Luke and Bassetti that called
this out in my comments. Uh, love Luke. A really smart, smart man. Um, but he was
actually also saying that becoming the most badass execution layer or like the biggest
pipe. Sometimes it's not even a question or a problem that
maps get to solve because they don't own the pipe like they're ultimately connecting to that
MTA. They are not their own MTA in the way that Salesforce Marketing Cloud and Adobe can be their
own. Right? Um, and so they're able to just throw money in compute at a, at a,
and queuing at a problem because they're, they own the MTA, whereas, um, other marketing automation
platforms that are built on top of an MTA and an aggregator, which for the record is almost every
other, it's most of its people. Yeah, the Adobe and Marketing Cloud are kind of the
two separate examples. They they build it from the ground up. But that's also because at that time,
probably one of the better options and only options. Yeah. I'll be really curious to see of the
more modern players who makes the first decision to become the actual
pipe? Yeah, it's definitely going to be interesting. And also how it's going to shift for like the
male guns, the in synch and bird spark posts. Like, what does that mean
for their business model? Okay, we've been getting extra nerdy, but now we're going to really dive
into talking tech. And I want to pressure test some things because, I mean, it's a pipe after all.
So let's get specific about what is actually running in production, because there's a
significant gap between the plain language example of what's actually deployed at scale, and
a marketing strategy based solely on a rules engine that can't support modern marketing teams
anymore. Conditions change faster than anyone can write. If then logic love me some handlebars or
liquid. So when you thought of this dumb pipe model and kind of just approach. Were you actually
talking about rules or about agents making judgment calls that rules can't anticipate? And if
it's the latter. Has anyone actually built that yet? Or is the whole category still promising? It
rules inside of a marketing automation platform. They take a few different forms, and you kind of
alluded to to a few of them, like nested segments, journey branches, filters, delays,
suppression logic, um, template logic, um, each new condition
adds another configuration layer. So like we're talking about inside of an application, uh, inside
of a user interface, like these are all things that you have to configure. And they're often all
in different places. And they also don't always neatly work against each other. Right. Um,
we've all definitely run into that one layer where it's like, oh, I had send time optimization
on, while I also had quiet hours and like the two just fought it out. Um, so I
think inside of the traditional layer, that's a decision tree and it
grows and grows and grows and it ultimately becomes harder to understand, harder to test
and maintain. Um, so I guess kind of the question that you were
asking is like, are these agents, are we talking about rules that humans are creating? I think it
has to be both. Um, I think there are a few platforms that are starting
to get it right, but we're still in a really nascent space here where like agents, I think
partially because they didn't always have all the context that was needed historically don't make
that great of decisions right now. I think that's going to change. But the core issue
is like, it's not that rules are useless. Um, it's
that encoding changing business context
and rules through a visual journey. Builder doesn't scale super
well. Because this has gotten more and more complex like it used to be. As simple as are they
opted in or not, but now people have. Once you get that, you have the appetite for the next thing.
It's like, all right, how are we doing? Like frequency management and making sure that we're
not annoying Peter with these messages. And it's the right amount of messages for Peter, right?
So I believe that the logic becomes a lot more maintainable when it's defined
upstream. And ultimately we're going to use agents for that. Um, because I don't think
that anyone actually likes setting up all of these rules inside of a user interface. Um,
so I'm the nerd that likes to. You are a unicorn. You
are completely solitary. Uh, I'm aware every generation of
martech platforms, whether it's a map or composable, launches with almost identical
language, unify the data, personalize the moment. Break down the silos. That's my
best radio voice I could ever attempt. And that's not very good. Is that convergence because the
pitch is actually true, or because it's what you say to close enterprise deals, regardless of
the architecture underneath it? It's probably a little bit of column, a little bit of column B, and
I'd also say like some of that marketing was used
because that was our best understanding at the time. Right. So I don't think it's necessarily a
matter of like, is this true or is it untrue? But I guess the question I would ask is like relative
to what is this true? So like unifying data is a good example. Are we talking about identity
resolution? Are we talking about simple merging of records. Are we talking about like deterministic
versus probabilistic? Um, and then I also asked the question like how manual is this or how
automated is this? And if it's automated and it makes a wrong decision, how do we fix that? Right.
Um, so marketing automation platforms, in my experience, are actually the worst at
unifying data. Um, they're not good at this. Um, especially not if you're
talking about identity resolution. So like for the most part, you're going to see deterministic
matching and a really rough process for merging profiles. If you failed to do that on the way in.
Right. Um, so that's probably the unifying of data side of
things. I guess the one other thing that I would say, and this relates back to the dumb pipe and
the whole data tax is like, how many of these platforms actually have useful
sunsetting policies that help you right size your contract. If only people
really thought about it. They don't. Right. And and, um, I also
think they shouldn't have to like that shouldn't be a part of it as like, oh, when when
they're no longer contactable and when they no longer engage with my app or my site or my
products, like, they should just naturally fall off. And then when they're due to come back, they come
back in. But I think the reality is, I don't know if this is a business decision, I don't
necessarily think this is like a nefarious thing that marketing automation platforms do, but they
don't make it easy. Um, so you know who does pardot. Oh, the recycle
bin. Oh, I love that. How much does Pardot pay you? Because you are there. Zero,
man. Literally $0. But it's similar to Exacttarget in that because they were the first
two real platforms, both for the B2C and the B2B side of the businesses.
They were figuring things out. And there's things Pardot offers that no other platform offers still,
and it makes zero sense to me. Yeah, but yeah, I love a recycle bin because it doesn't count
against your contact, your marketable contacts, and it's great. Yeah, like there are platforms that
start to get into this. Um, like I think for certain types of businesses, the MOU
model makes some sense. Um, I'm still not a fan. And as you can well tell, like,
I'm not a fan of charging people to use a copy of their data. Um, so
unifying the data like that to me is more sales pitch than anything. Personalizing the moment.
Uh, marketing automation platforms are probably a little bit more solid in this area, right?
Like, I think everyone's use. I would hope so. Right? Like, I think everyone's pretty used to marketing
automation platforms being able to, uh, personalize content for you to be able to preview that
personalization and make sure that, uh, the rules that you created to call back to the
rule creation inside of maps is, is correct. Um, I think they do that with different levels
of ease. Right? Like, it's very easy to do that in some platforms. They have the tooling that makes
it easy. It's more difficult in others. So personalization, they absolutely get, right. Um, the
last one you mentioned was breaking down data silos, which is like the funniest one. Um, because
by definition, most marketing automation platforms are creating silos. Um,
not explicitly stated, but they are in the business of creating copies of data and creating
silos, not breaking them down. Yeah, it's very reminiscent of the original CDP debate, because I
remember the first time I was introduced to a CDP ten years ago, I was like, why? I don't see the
reason or point to pay for this, right? And that was also an argument that I historically
had is like the traditional CDP. The packaged CDP, if you will, is a product
without a buyer persona. Data engineering teams didn't like working in them.
Marketers were also pretty scared of working in them. Um, and they in both
cases prevented those consumers for working in the platforms that they wanted
to. It was a bandaid solution to a different type and different era of data silos.
And instead, let's create a different data silo just for this use case. Right. And a lot
of it was focused on like, all right, how do we capture more real time data from your site or
your app? Um, but now there are other tools for that. Um, and, and it was kind of pre
full adoption of a warehouse tool. Oh, yeah. And so where were you going to get your streaming data.
Where were you going to be able to get these tracking events. So yeah, to put them involved to
put a bow on the data silos. Like I guess you could make a point that you're breaking down
silos if like, you're consolidating multiple solutions into a single map. Like if you're
using Attentive and Salesforce today and you're just moving everything over to a Bres or an
iterable, like you could maybe claim that you're reducing a data silo there. Um, that's true. But
like at the end of the day, data silos doesn't hold much weight because they're taking a partial
copy of your data. And then on top of that, they're also creating data in your
engagement events. And you have to get that back to your warehouse or whatever your central source
of truth is to be used for analytics and decision making and all of that. So like in terms of
unifying data, not so great personalization. Yeah. Good. Breaking down silos. Absolutely not. Nope.
Yeah. And the one counterargument is if you're at a company where your data is inaccessible, it is.
There's a great wall of China. No one can pass. And your marketing team
needs to do something and needs to market and contact customers in some way, shape
or form. That is the rare instance where, yeah, that data silo is your lifeline. It's your only option.
So it's a tricky thing, but it's all to your point at the very beginning. Like you have to have so
many clarifying questions to know what the right solution is for your current state of the
business, and then your future state of where you want the business to go. Next time we go deeper
into what it actually takes to close the gap, the economics nobody's pricing in yet. And Peter names
names on which maps are adapting and which ones are refusing to stick around.