Martech Therapy: CDP strategy and agentic AI

Martech strategy

What your Martech vendor's filings tell you before you renew

Your vendor's annual report is free, public and full of things that change a renewal conversation. Including the one date nobody tells you they care about.

A bank buying a customer data platform has somebody, somewhere in the building, whose job is to check whether the company selling it can afford to still be there in three years.

A retailer buying the same platform on the same three-year term usually has no one doing that at all.

Same product, same paper, same length of commitment.

The difference is which industry you happen to work in, and whether anyone thought to open a document that's been sitting in public the whole time.

Does anybody actually do this?

In one corner of the market, yes, and

Martech readiness

Use cases should lead. Lead with what?

Everyone agrees use cases should lead a platform decision. I agreed in print, then reached for the requirements template anyway. Here is what was missing.

Some time ago I wrote about why skipping the RFP isn't always an option. The piece opened on a comment Jay Calavas from Tealium had left under one of my LinkedIn posts:

"Skip the RFP and focus on use cases and measurable outcomes!"

I said he was right. Use cases should lead. I meant it, and would say it again tomorrow.

Then I sat down in the next selection and couldn't act on it.

I reached for the requirements template, because, look, it was there and the clock was running. The plan was that the use cases would

What real time actually costs

Real time is several different speeds, with very different bills. What each one is for, and the cost nobody likes to mention.

Sponsored by Databricks. Databricks commissioned and funded this guide and reviewed it before publication for factual accuracy. They did not get to change the conclusions.
Full disclosure at the foot of this post.

Somebody says "real time" in a meeting and the room nods.

It is the most agreeable phrase in martech, because everybody in the room assents to something slightly different and nobody finds out until the CFO starts looking for you.

I wrote about the architecture of this in July, in real-time from a warehouse:

How quickly context goes stale, where the work should sit, and

AI in Martech

Natural-language audience builders: my advice for vendors

Natural-language data interfaces let marketers skip the BI ticket, and the data engineer with it. Some advice for the product teams building them.

I talk to a lot of martech product teams. Part of my advisory work is getting an early look at features before they go GA, sometimes months ahead… well, weeks, more often, and giving feedback while it can still change something. Most of those previews have little in common. Lately one pattern keeps showing up, across companies that don't know they're building the same thing.

A marketer types a question in plain language. An LLM queries the warehouse and comes back with insights, an audience, and sometimes a full activation plan. No SQL, no BI ticket, no waiting

Martech contracting

What is a forward deployed engineer, and who is paying for the one in your team?

Vendors are embedding their own engineers to build your stack. One Martech company's accounts show the implementation running at a gross margin of 0.08%.

At the Martech World Forum in London this month, Nate Wardwell from Hightouch talked about forward-deployed engineers and described them as a service Hightouch provides that is helping its customers. I wrote it up in my recap and left that question hanging.

Isn't that what independent consultants do?

I went and looked properly.

What I found wasn't really about job titles.

What a forward-deployed engineer actually is

In short, someone the vendor employs, who sits inside your team and builds the thing. Not a trainer, not a support contact, not a consultant who writes a plan

Customer Data Platform

How to write a data request someone can act on

A request rarely gets refused. It gets built, exactly as written. Here are the six decisions to name before you send one. One page, free, no forms.

Sponsored by Databricks. Databricks commissioned and funded this guide and reviewed it before publication for factual accuracy. They did not get to change the conclusions.
Full disclosure at the foot of this post.

A badly written data request rarely gets refused. It gets built.

That is the part people miss when they complain about the queue. You ask for lapsed customers, three weeks later you get lapsed customers, and the team who built it did exactly what the message said. Then the campaign underperforms and nobody can point at the mistake, because there was not one.

Lapsed meant a

Latest Posts

Does anybody actually do this?

What a forward-deployed engineer actually is