The data strategy sequencing error: why you're starting in the wrong place

Almost every company I speak to has already invested in data: an analytics platform, a data warehouse, perhaps an initial AI pilot project. Yet when I ask, "What does your data allow you to do today that you couldn't do a year ago?" the answer is often vague.

The tools are there, but the question they're supposed to answer isn't always there.

This is a typical sequencing error: you start by thinking about the tools and the people who will use them. You start with the how, not the why.

Check the fit between data and product strategy

Before wondering what tools are needed or how to structure your data team, there's a simpler and often obvious question to ask: is the data you have (or could have) consistent with the product strategy you're pursuing?

Let me give you an example that often comes up in B2B software. A company decides to shift its product model toward expansion - upselling, cross-selling to existing customers - after years focused on acquisition. The strategy is clear, and the market justifies it. But the data collected so far, designed to measure conversion and first purchase, says nothing about usage, feature adoption, and signs of expansion.

The gap here is clear: the existing data answers a different question than the one the new strategy is asking.

And this is precisely the first check to make, before any other: not optimizing data collection or data quality, but understanding whether you're measuring the right things for the direction the product is actually taking.

Why your data strategy shouldn't start with tools

I know, it's tempting to approach the data topic from a tactical level: which platform, which team, which expertise to hire immediately. And it's also the most common way data projects fail: infrastructure is built before clarifying the decisions it's supposed to be used for.

The starting point that works is another, and I say this with the same sincerity with which we ask our clients: what would data allow you to do if you had it in the right format?

Not "what dashboard do you want," but "what decision do you make today based on instinct that you'd like to make with clarity?"

It's a strategic question, not a technical one. The most helpful answers I receive at this stage rarely mention a tool. They usually describe a recurring decision - for example, prioritizing the roadmap, understanding where customers are being lost, calibrating pricing - that is currently made without solid evidence.

Only after clarifying this does it make sense to delve into the infrastructure, team, and tool levels.

First the what and why. Then the how.

Build your data strategy gradually, not in one leap

There's a widespread expectation - fueled more by fomo than facts - that AI is a shortcut to bypassing the work of data consolidation.

It isn't, and there are no shortcuts here. Any AI initiative, be it a predictive model, an internal assistant, or a recommendation engine, relies on consistent, traceable data, linked to the decisions it must inform. If that foundation doesn't exist, AI doesn't solve the problem. It automates it, only making it harder to identify.

This is why the useful work isn't choosing between "first we consolidate the data" and "first we adopt AI." It's identifying the right sequence of problems to solve, in the right order for where the company actually is, not where it wants to be.

Sometimes the priority is to organize a handful of metrics shared across functions. Other times, that foundation is already there, and the next problem is organizational: decisions that remain slow even when the right data is already there, available. These are different interventions and should be addressed sequentially, not all at once, hoping that AI will fill in the gaps.