You need a product ICP to minimise churn

In marketing, we talk about ICPs - Ideal Customer Profiles - all the time. They help define who we want to sell to by describing the individuals and companies that are most likely to benefit from our product or service. The closer someone matches that profile, the more likely we are to prioritise our time and effort on them. Done properly, it’s a way to grow faster and more efficiently, because the product aligns well with their needs.

Recently, I’ve been thinking that we need something similar in product - an ICP for product development.

As a CTO in a SaaS company, one of my constant priorities is keeping the product roadmap aligned with what our customers actually need. That means having regular conversations with customers. In theory, it sounds simple: talk to users, ask good questions, find the most valuable things to build next. But in practice, it’s not that straightforward.

Some conversations are surface-level. You go through the motions, ask the usual questions - what’s working, what’s not, what’s missing - but the answers don’t always give you clarity. Feedback can be vague, tactical, or sometimes completely contradictory, especially when you're talking to different regions or roles within the same organisation.

But every now and then, you talk to someone who really gets it. Someone who not only uses your product, but also sees its potential and can articulate meaningful ideas for where it should go next. When that happens, it’s like finding gold. You leave the call energised, full of ideas, and importantly with a clear sense of direction.

Finding “John”

One of those conversations stood out recently. Let’s call him John.

John is Head of AI at a large software company. He’s hands-on with our product and has been since they became a customer a few years ago. The first time I heard about him was from our customer success team. They’d tell me how John was solving business problems in ways we hadn’t even imagined pushing the boundaries of the product, creatively stitching it into their workflow, and generally doing things that made me think, “Why didn’t we build for that use case?”

When I finally got on a call with John to talk roadmap, it was immediately different from the usual product feedback session. He wasn’t just listing feature requests or small annoyances. He had ideas that were strategic but grounded. Real problems, clearly defined, and framed in a way that made sense for both their business and our product’s evolution.

For example, he proposed adding simulation capabilities to support “what-if” scenarios - things like forecasting the impact of changing the owner of a sales opportunity, or understanding how launching a new product line would affect their pipeline. It wasn’t a tweak to a field label or a colour on a chart. It also wasn’t a moonshot about AI reading market signals and telling reps what to do. It sat in that sweet spot: ambitious, practical, and directly tied to business value.

He had two or three other ideas in the same league. Each one had the potential to unlock serious value, not just for his company, but for others too. And the best part was that we could immediately see how they’d fit into the roadmap.

That one conversation was worth more than two months of standard customer calls.

Not All Feedback is Equal

I want to be clear: I’m not saying the other conversations aren’t useful. They are. You always learn something when you talk to customers. But not all feedback carries the same weight when it comes to shaping product direction. Some people give you a laundry list of minor issues or one-off use cases. Others dream big but can’t tie it back to real workflows.

What I’ve realised is that the person you’re talking to matters just as much, if not more, than the questions you ask.

This is where the idea of a Product ICP comes in.

The Product ICP

If you want to get meaningful input on your product roadmap, you need to be talking to the right kind of user. And in my experience, they tend to have a particular mix of traits. Here’s what I look for:

  • Key influencer or decision-maker when it comes to renewal or upsell

  • Hands-on with the product or at least closely involved in how it’s used

  • Understands the value the product delivers to the business

  • Responsible for solving business problems, not just technical ones

  • Has a team who use the product and is in touch with their needs

Not everyone with these traits will give you great feedback, but you’re far more likely to get something valuable from them than from someone who just inherited the tool or only sees it through the lens of a single dashboard.

If you’re just scheduling customer calls at random or picking whoever replies first you might be wasting the opportunity.

Finding Your Own John

So how do you find these people?

In my case, it was through internal conversations with the CS team. They had visibility into who was doing interesting things and who really cared about the product. That was my clue to reach out.

It might be someone in a technical leadership role. It might be a power user who’s built internal tools or automations on top of your product. It might even be a manager who constantly asks tough questions about how the product could do more.

Once you find someone like that, make the time. Prepare properly. Ask real questions. And don’t waste the opportunity by focusing on minor UX tweaks.

When you hit the right conversation, it’s obvious. The energy shifts. The ideas are clearer. And you walk away with input that actually helps shape where the product should go next.

Final Thoughts

John changed how I think about customer feedback.

We still talk to a range of users, and we still value every piece of input. But when it comes to prioritising roadmap decisions, I now put more weight behind conversations like the one I had with him.

If you’re building a product and trying to reduce churn, increase engagement, or drive upsell, then you need to start thinking about your Product ICP the same way your marketing team thinks about theirs.

Find your own John. And treat them well.

Gareth Davies

Gareth is an AI researcher and technology consultant specialising in time series analysis, forecasting and deep learning for commercial applications.

https://www.neuralaspect.com
Previous
Previous

How AI Actually Works

Next
Next

How I make and deliver a roadmap