For the past month, I’ve struggled to log in successfully to the digital version of a famous print publication. Please indulge me as I convert a personal hassle into some larger advice about the future of MarTech. In short, identity snafus are merely the most visible symptom of a much larger infrastructure challenge—and AI is about to make your infrastructure choices far more consequential.

Let’s begin with a famous aphorism (popularized originally in 2005 by Tim Bray ) from the late computer science professor Phil Karlton, There are only two hard things in Computer Science: cache invalidation and naming things.

We could coin a narrower offshoot: There are only two hard problems in MarTech: identity resolution and naming things.

On the naming-things problem, check out this podcast episode with guest Stephanie Lemieux on the future of taxonomies. Today, let’s talk about identity resolution—often acronymed to “IDR”—as a proxy for larger infrastructure challenges that loom in your increasingly AI-inflected stack.

My Login Problem

At some point back I stumbled into creating two trial accounts for Publication X, part of a large conglomerate, and it seems I registered an account with at least one other of their publications. You might recognize the circumstances: I created all of them at different times, likely within different browsers on separate devices. I ponied up for a paid version of Publication X two years ago, but the other accounts still lurked underground like cicadas in the loamy dark recesses of the publisher’s customer data stores.

After a confusing array of renewal messages—some no doubt for “win back” campaigns—I thought I renewed my paid subscription, but turns out I may have actually created a new record. This has messed up my access across devices in myriad ways I won’t bore you with. It’s a minor annoyance that I’ll try to resolve by calling customer support, but something makes me think that the same problem could recur.

The MarTech advisor in me has some empathy for this publisher. In recent years they seem to have been building a conglomerate-wide identity management platform, which makes sense. But they will inevitably struggle to execute across diverse brands with different legacy data and business models. I’ll guess that in many subterranean burrows of their data ecosystem, my name shows up as a predicate field in rows where the initial offer or transaction was the primary key. It takes a lot of time and money to un-pretzel all of that.

Your IDR Problem

This publishing company is not alone. I don’t know a single large enterprise that isn’t struggling with some dimension of IDR, although those firms that have worked on it the longest seem to have the best handle today.

It certainly comes up a lot in advisory sessions we’ve had at Real Story Group with some of our corporate members recently. A short selection of particularly thorny IDR topics among our members have included:

When a B2B enterprise markets more to end-users or consumers When a B2C enterprise has its most comprehensive consumer data tied up in an external Loyalty datastore (yikes!) When a multibrand firm with millions of recorded purchase transactions tries to reconcile individual identities (see pretzel discussion above) When an enterprise needs to start tracking consumer-agent bot activity…

These present to MarTech leaders as customer-data puzzles, yet underneath they signal foundational infrastructure demands. Identity just happens to be where they surface first.

The Broader Infrastructure Mandate

MarTech seer Scott Brinker recently posted about what he calls the coming Infrastructure Wars . He makes some important points about an evolving competitive context around various tiers of vendors. I would just add that the infrastructure wars have been going on for some time, as MarTech suite vendors jockeyed over the past decade to become your data / content / personalization store of record. In most enterprises, Marketing and MarTech leaders like you successfully fought the good fight against this suite pitch, in order to sustain a composable stack.

Figure: Your MarTech stack is evolving lower and lower into infrastructure

But in the meantime, a different infrastructure story is playing out. IDR is just one issue among many that recommends MarTech leaders like you to become more active participants in your enterprise infrastructure around data, content, and decisioning.

This is a tough ask I'm making. You're already getting sandwiched between executive pressure for results coupled with uneven feasibility and ROI among your AI pilots. Yet, a big part of the solution lies in getting closer to your Data and IT peers. They too are getting sandwiched, but you can all become more worthy allies in an AI future that will see changing roles and some important “build” (versus "buy") projects.

This kind of alignment is not always easy. Within some of RSG’s corporate members, things like IDR fall to the MarTech team to build out, rather than the enterprise Data organization. For some of you, AI will need to run a similar course. Don’t underestimate the technical complexity, but at the same time, the wins from centrally managing essential services like context and orchestration can be huge.

Some Final Advice

Let’s pause to acknowledge that your firm can deploy the highest-quality identity and other customer data, and still create brand-eroding customer interactions due to inattention to actual experience at the glass. But likewise, it’s impossible to create effective customer experiences without a solid base of data and content. The best UX in the world was still not going to log me into that magazine article I wanted to read.

Even as stacks mature, the most pressing questions in MarTech revolve around architecture. The advent of AI only magnifies those choices. As you weigh your alternatives, know that RSG can offer experienced, real-world advice. Reach out any time .

P.S., to circle back where I started this post, cache invalidation is going to get even harder in a world of AI Agents and complicated, just-in-time experience assembly. It’s yet another infrastructure problem! So, best wishes for further collaboration with your Data and IT teams. Let me know if I can help…