systems of context and systems of truth – Chief Marketing Technologist

For years, martech was divided into methods of document and methods of engagement.

Techniques of document saved the “grasp” variations of knowledge. CRMs and CDPs for buyer knowledge. PIM for product knowledge. DAM for model property. ERP for stock knowledge. (I do know, there’s a ton of acronyms on this piece, nearly to the purpose of parody. Click on on the picture above for a larger-scale model to learn the Acronymn Decoder Ring on the best.)

Techniques of engagement interacted with prospects, both straight with instruments corresponding to MAP-delivered e mail or DXP-delivered internet experiences or not directly with instruments corresponding to SFA supporting gross sales groups who engaged with prospects.

In actuality although, this division wasn’t fairly as clear. As an illustration, CRMs and CDPs usually blended knowledge administration and buyer engagement performance. However as a method of speaking about totally different martech classes and their roles in a tech stack, it sufficed.

However I don’t suppose it’s the best psychological mannequin anymore.

With the AI revolution sweeping all the pieces in advertising and know-how, I consider a greater framing is methods of context and methods of fact.

Now, squinting at my diagram you may ask, “Isn’t this simply methods of engagement and methods of document with totally different labels?” They’re related, sure. Each roughly delineate a layer of knowledge and a layer of companies. However right here’s why they’re totally different.

First, on the knowledge layer, methods of document traditionally mixed obligations for each storage and arbitration of knowledge. Every system of document had its personal database that was tightly coupled with enterprise logic to find out what might be written to or learn from that database and in what format.

Within the mannequin I’ve sketched right here, the methods of fact part of the stack acknowledges that these issues of storage and arbitration have now been separated, due to the rise of the cloud knowledge warehouse/lakehouse.

Cloud knowledge warehouses/lakehouses retailer and distribute knowledge throughout the group. However the very power of this common knowledge layer — that it might home any and all knowledge flowing throughout the group — can be its weak point. Knowledge flooding via it isn’t essentially standardized or rationalized throughout all its varied sources and locations. Battle and competition can run rampant.

You continue to want software program to arbitrate what’s right and canonical knowledge, governing the way it will get written or learn, in what format, validated with commonplace definitions and related enterprise logic. With one thing as necessary and complicated as grasp buyer data — what’s historically saved in a CRM — such arbitration and governance is non-trivial and mission-critical.

That is why many “basic” martech methods of document — CRM, MDM, PIM, DAM, and so on. — nonetheless play an necessary function within the cloud knowledge warehouse/lakehouse period: they continue to be the arbiters of fact for knowledge inside their area. Even when their knowledge is more and more saved in an impartial layer additional down.

Will there be a brand new technology of such knowledge arbiter platforms? Or will the present main platforms of at present evolve to adapt to this new setting? In all probability each.

These basic martech knowledge platforms additionally present helpful contextualization of knowledge. As an illustration, combining an inventory of shoppers with a set of promoting viewers segments creates a context for that buyer knowledge inside a selected advertising marketing campaign. That is exemplified with composable CDPs. They work straight with the info saved in a cloud knowledge warehouse, however they arrange and handle that knowledge for all kinds of various contexts through which entrepreneurs need to use it.

In actual fact, a composable CDP is arguably extra system of context than system of fact.

(If the phrase “fact” bothers you for philosophical causes, and also you’d want we alter it to one thing else, sorry, we Kant.)

We by no means achieved a real single system of fact (SSOT). Turns on the market are simply too many domain-specific knowledge truths. However with a common knowledge layer on backside and domain-specific knowledge governance platforms on high, we now have many methods — plural — of fact.

So what makes methods of context totally different from methods of engagement?

Techniques of engagement have been comparatively “mounted” within the context they supplied. As an illustration, with MAP and CEP platforms, you discovered how you can use them, usually adapting the way in which you’re employed to their constructions and processes. While you constructed a web site on a DXP, the expertise prospects obtained was the context you had thoughts for them whenever you designed it. It’s not likely their context. It’s the context you suppose they’ve.

With the rise of AI brokers, each employee-facing and customer-facing, context is being created extra dynamically. Many AI brokers might be spun up, every tailor-made to a selected job or workflow for an worker or hyper-personalized for a selected buyer’s expertise.

Techniques of context differ from methods of engagement quantitatively — more and more, there are extra AI brokers proliferating throughout the stack than conventional SaaS platforms — and qualitatively as a result of they’re purpose-fit for far more specialised contexts.

The acute incarnation of that is particular person AI brokers that create software program experiences on-the-fly for every worker or buyer, for no matter job-to-be-done they need accomplished at that second. That is what we described as the brand new “hypertail” of martech software program — in aggregrate, billions of software program apps created on-demand by AI — in our Martech for 2025 report a few months in the past.

Dynamically-generated customer-facing AI brokers, which we’ll name concierge AI brokers, will ship far more contextually related experiences to these customers. They’ll hear to precisely what the client needs, and knowledgeable by all present buyer and firm knowledge in our methods of fact, ship precisely the content material and companies the client needs in that exchage.

As a substitute of a monologue, the place a model serves a contextual expertise based mostly by itself definition of the client’s journey and what it thinks the client needs, concierge AI brokers will have interaction consumers in a real dialogue to grasp and serve their precise context.

Somewhat farther out on the horizon — however possibly not that far — consumers will use their very own AI brokers to work together with our methods of context. These aren’t customer-facing AI brokers that sellers management. These are customer-owned AI brokers that they management. They may inherently form the expertise to the context of the client.

Two extra factors:

First, within the stack schematic on the high of this submit, I included an entire bunch of various martech product acronyms. I’m not saying that each stack wants all of those. Except you might be at a big enterprise, you in all probability solely have, want, or need a subset. I solely included a bunch to point out the place I believed these totally different classes match on this structure. It’s also certainly not complete.

Second, illustrating this as a stack — parts neatly packed collectively, stacked on high of one another like completely becoming Lego blocks — isn’t a very correct illustration of actuality.

The stack view is simpler to grasp — a minimum of I believe so, provided that’s how we’ve thought of martech stacks for therefore lengthy.

However a extra correct illustration can be a graph view. All of those totally different merchandise and platforms, and all these totally different apps, brokers, and automations, are all nodes within the cloud that join with any of the others. I nonetheless consider methods of fact on the middle of this cloud, surrounded by many methods of context.

In abstract, sure, we nonetheless have data, and we nonetheless have engagement.

However the defining attribute of the martech “stack” in an AI world goes to be context and the fact it’s wrapped round.

P.S. Talking of martech stacks, we hope you’ll take into account sending in a slide illustration of yours for our 2025 Stackie Awards. It’s significantly one of the enjoyable awards packages in all of selling — admittedly in my extremely biased opinion. However there’s no price to enter, and in reality, we donate $100 to UNICEF for each legit entry, as much as $10,000 whole. Deadline to enter is April 4.