Segmentation is often the most over-promised — and under-delivered — research tool. Here’s why.

Segmentation is one of the most beguiling — and challenging — types of research. Not because of the math, but because oversized expectations often collide with ambiguous direction…

“We want to segment our customers.”

That’s the typical starting point for so many projects. But your first follow-up question absolutely must be:

“For what purpose?”

Challenge #1: Application

When I first started doing segmentation, I hit a lot of walls. We’d go through the whole process — segment the customer base, build out fancy personas, give them clever names, craft beautiful decks explaining who they were and why they mattered.

And then… crickets.

Whenever it came time to make an actual decision — messaging, product, targeting — everyone would quietly veer away from what the segmentation supposedly told us. I couldn’t figure out why.

The hard truth: we never really designed the segmentation for anything specific.
We weren’t clear on the intended use case.

Was it for media buying? For product roadmap decisions? For creative briefs? For targeting CRM? For internal culture-building?

If you don’t define this up front, no matter how brilliant your segments are, they’ll end up collecting dust.

Challenge #2: Replication

Even if you do nail the intended purpose, if you can’t replicate the segmentation consistently — across studies, trackers, new waves of data — it quickly becomes irrelevant. People stop trusting it.

Many segmentations cast too wide a net — pulling in everything from demographics to category behaviors, spend, time, and perceptions. You often end up with segments built from 30+ variables. While that may sound comprehensive, it creates a bulky, unwieldy typing tool that won’t fit into most future studies. The result? The segmentation gets left on the cutting-room floor — and your carefully crafted segments stale like forgotten bread.

What does all of this tell us, grasshopper?

There is no one perfect segmentation.

At my last role running research, we ended up building five segmentations.
That sounds crazy — but each was tuned to a specific internal audience and purpose.

• The media team had one.
• The product team had another.
• The brand team had theirs.
• Leadership had one view.
• CRM had its own operational view.

No single team needed to master all five. They each needed to understand the one that mattered to their decisions. The onus fell on us, the insights team, to master understanding and to manage the connective tissue behind the scenes — to ensure the models were coherent, that the architecture made sense, that the organization didn’t fragment into chaos.

Think of it like a product architecture — as a skilled product manager, you wouldn’t launch a product line of five products with no shared foundation. Segmentation is no different.

3 Key Questions:

So the next time someone says, “We want to segment our customers,” here’s the checklist:

1️⃣ For what purpose?
2️⃣ How will you replicate it?
3️⃣ Are you okay with multiple segmentations — and do you have a plan to manage them?

Answer those well, and segmentation becomes a tool that actually drives decisions, not just fills PowerPoint decks.

Next Round

In the next round, I’ll share my approach for building segmentations that stay fresh and deliver impact.

If you’d like to dive deeper into segmentation — or explore how I can help you truly understand your customers — I’d love to connect.

Ready to wrestle your segmentation into shape? Let’s talk.