A good while ago now, I wrote that Adobe Journey Optimizer is not a marketing automation platform.
I presumed this was a fairly mundane observation. Unfortunately, the martech industry has spent decades convincing itself that every new platform is the long-awaited replacement for all previous platforms. Mostly because people tend to become emotionally attached to categories. We like things to replace other things. We prefer delusional evolutionary stories in which one species elegantly replaces another. This, however, does not reflect reality. Homo sapiens, homo neanderthalensis, and Denisovans all coexisted. There are no neat little lines to distinguish one from another.
Enterprise software behaves in much the same fashion. New layers arrive. The old layers never quite disappear. Systems integrate and merge in myriad ways that would challenge the most liberal interpretations of appropriate (in)breeding…
After spending the better part of the last year helping enterprises evaluate and implement AJO B2B, I’ve become convinced that the most interesting challenges have very little to do with the technology itself.
The technology is mostly ready. Your organization probably isn’t.
Most implementations begin in an atmosphere of optimism. There are discussions about buying groups, account journeys, personalized engagement, customer signals, and AI-assisted everything. Architects draw diagrams. Vendors demonstrate capabilities via PowerPoint but never in the platform itself. Executive sponsors talk about transformation while wearing a hero’s cape. People that get paid more than me go and eat steaks and high five.
Then someone asks how account hierarchies work.
Not how they’re supposed to work.
How they actually work.
And there is a notable dearth of high fives at this point.
Sales has a hierarchy. Finance has a hierarchy. The CRM has what appears to be a hierarchy, though nobody can remember when it was last reviewed. Somewhere, somebody has a spreadsheet that everyone agrees is authoritative until a different spreadsheet suddenly appears. Snowflake and Databricks continue to blindly serve up whatever the last guy asked for.
This is not unique.
In fact, the longer these projects run, the more frequently the same pattern emerges. What begins as a technology implementation gradually transforms into an investigation of assumptions everyone was certain had already been addressed sometime during a budget cycle long forgotten by history.
A single person somehow exists as multiple people, or multiple profiles at least. And profiles aren’t people. A single email address frequently belongs to multiple profiles. And for some reason default behavior is to just send it to them all.
Joker: “Have you shipped a million of those things?”
Scientist lookin guy: “Yes sir!”
Joker: “SHIP ‘EM ALL! WE’RE GONNA TAKE ‘EM OUT A WHOLE NEW DOOR!!!”
Ok then, how about lead routing. Lead routing operates according to rules that are critical to the business and understood by no one currently employed. Oh, and AJO still doesn’t do that. At least not without an insane amount of effort.
Lead scoring is simultaneously important, poorly documented, and owned by a committee that no longer meets. It, as well, is not native to AJO in a meaningful sense. You want it? You build it.
Consent is scattered across several systems, each of which considers itself authoritative. Customer ownership depends largely on who is being asked and whether legal, sales, marketing, or finance arrived at the meeting first. Did you have a meeting about arbitrating the 5 separate “golden” records? You did not.
What makes this particularly entertaining is that many practitioners arrive expecting Adobe Journey Optimizer B2B to solve these problems in the same way previous marketing automation platforms appeared to.
In reality, AJO often treats these not as platform features, but as implementation decisions. Sometimes it provides flexibility. Sometimes it provides building blocks. Occasionally it provides a strongly worded suggestion. More often than not, however, it hands the responsibility back to the organization and waits.
The software is remarkably calm about this.
It does not particularly care how many definitions of a customer your org maintains. It simply requires one when it’s time to build an audience. It does not object to multiple account hierarchies. It would merely appreciate knowing which one is real, or at least real enough to activate the journey. It has no philosophical position on lead scoring, lead routing, consent ownership, identity management, or attribution models. It takes for granted that the adults have already discussed these matters.
This assumption is frequently always optimistic.
For years, marketing automation platforms quietly absorbed much of this ambiguity. Leads were routed. Scores were calculated. Programs accumulated members. Email went out. Reports appeared. Nobody particularly enjoyed maintaining these systems, but they possessed the useful quality of already existing. An entire generation of marketers inherited operational machinery that felt as natural and permanent as indoor plumbing.
Then somebody starts implementing AJO B2B and discovers the plumbing is in fact connected to six other buildings, three acquisitions, a CRM, a data warehouse, a lakehouse, and for some reason there’s a urinal in the basement. Which is… genius?
What I’ve come to appreciate about Adobe Journey Optimizer B2B is that it doesn’t create most of these problems.
It simply stops hiding them.
Many experienced marketing automation nerds arrive expecting familiar concepts to be native platform behaviors. One email address should receive one message. Right? One person should be one profile, after all isn’t that the whole point of the CDP? An audience of ten thousand should produce ten thousand journey entrants. C’est vrai? Lead scoring should exist. Lead routing should exist. Attribution should exist. Operational reporting should answer operational questions. The marketing database should know who the customers are.
Reasonable hypotheses.
But uh, nope.
Or perhaps more accurately, many of them are no longer defaults the platform is willing to make on your behalf.
This distinction turns out to matter quite a lot.
Because once a platform stops deciding these things for you, somebody else has to. The bummer is that it ends up being a collection of stakeholders who have spent the last decade trying to avoid this discussion altogether.
This is where many implementation teams begin talking less about technology and more about governance, ownership, operating models, data stewardship, lifecycle definitions, buying group composition, attribution methodology, and all the other topics capable of draining the enthusiasm from a conference room at record speed. And us MarTech folks figure out that we’re stuck waiting for them all to make some decisions. And wait we shall, and wait we do.
Incidentally, a colleague recently said something that I took as a complement.
“You are always smiling so much when you’re delivering the worst news…”
Oddly enough though, this is often the moment the project starts creating real value.
Not because bad news is fun.
It is not.
Not because the answers are obvious.
They are not.
But because transformation was never really about a new platform. That bit arrives on day one. The difficult part is discovering that half the business operates on hunches, a quarter operates on workarounds, and the remaining quarter is barely holding it together because man, we all got SOMETHING going on.
AJO, in fairness, never claimed it would solve any of that.
A salesperson, nevertheless, might have.
The interesting thing about AJO B2B is that it remains surprisingly bad at solving problems you’ve spent fifteen years assuming were platform features, while being exceptionally good at orchestrating experiences once those problems have already been solved.
Unfortunately, those are not solved simultaneously.
Most of the friction I see hasn’t come from missing features. It’s come from discovering that what many businesses considered platform functionality was actually a collection of old decisions, undocumented processes, and tribal knowledge. That and maybe a lack of transparency from my dear friends that make this thing.
The software was never the hard part.
It was the humans all along. The monsters are just people in masks. Or maybe “We have met the enemy and he is us.”
If you’d like help sorting out which assumptions deserve to survive the journey, I would be happy to compare notes, preferably over steaks paired with an overpriced French wine whose left-bank terroir can finally justify all those discovery sessions and governance workshops.