---
title: "Kim Beyns on how NGroup tests a decision before it commits."
dek: "A few minutes with the CEO of NGroup — NRJ, Nostalgie, Chérie — on using persona panels as an everyday decision tool. His words, and what I think they mean for any media leader."
type: Essay
author: Samuel Zniber
datePublished: 2026-05-21
dateModified: 2026-05-21
wordCount: 620
readingTime: 4 min
url: https://www.samuelzniber.com/insights/kim-beyns-ngroup-persona-panels.html
canonical: https://www.samuelzniber.com/insights/kim-beyns-ngroup-persona-panels.html
section: Insights
tags:
  - persona panels
  - predictive audience modelling
  - decision intelligence
  - radio programming
  - NGroup
mentions:
  - https://www.musicdatak.com
  - https://mediadatak.com
video: https://www.samuelzniber.com/kim-beyns-ngroup.mp4
heroImage: /kim-beyns-ngroup-poster.jpg
heroImageAlt: "Kim Beyns, CEO of NGroup, in conversation about persona panels"
relatedSlugs:
  - perspective
  - case-studies
---

# Kim Beyns on how NGroup tests a decision before it commits.

> A few minutes with the CEO of NGroup — NRJ, Nostalgie, Chérie — on using persona panels as an everyday decision tool. His words, and what I think they mean for any media leader.

**Key takeaway** — Kim Beyns runs NGroup — NRJ, Nostalgie, Chérie. In this conversation he describes using a panel of 100 modelled listener personas as an everyday decision tool: querying it for summer programming, narrowing it to the under-35s, and pre-testing ad creative and music before the budget is spent. The framing to take away is his — it is a complement to traditional research, not a replacement.

▶ Watch the conversation (2 min 46 s): https://www.samuelzniber.com/kim-beyns-ngroup.mp4

---

Every media leader I work with is carrying the same quiet problem: the decisions keep getting bigger, and the time to make them keeps getting shorter. So when one of them has actually closed part of that gap — and is candid, on camera, about how — I pay attention.

The conversation in this piece is a few minutes with Kim Beyns, who runs [NGroup](https://www.ngroup.be) in Belgium — the group behind [NRJ](https://www.nrj.be), [Nostalgie](https://www.nostalgie.be) and Chérie. I should be straight about my interest here: predictive audience modelling is what we built [MediaDatak](https://mediadatak.com) and [MusicDatak](https://www.musicdatak.com) to do, and NGroup works with us. That is exactly why I wanted Kim to describe it himself, from the CEO's chair, in his words rather than mine. Here is what stayed with me.

## A panel of a hundred listeners, open at any hour

NGroup built a panel of 100 modelled listener personas, drawn from real, public data about their market. What struck me is not that it is clever — it is that it is *available*. The panel does not keep office hours. When a question comes up, the team asks it and has an answer in seconds.

Last week the question was the summer. Should the morning show hold its shape through July and August? Should NGroup put on concerts? They put it to the panel — and then did something traditional research rarely lets you do on a Tuesday: they asked the same questions of the under-35s alone.

> "It's so rich — you get answers within seconds."

## The sentence I would underline

If you take one thing from the conversation, take this — and notice that it is the CEO saying it, not a vendor:

> "It's important to say this kind of research doesn't replace traditional research. … This is an extra research, and this is so rich."
> — Kim Beyns, CEO of NGroup

He is right. NGroup still runs its annual focus group — seven real people, its own listeners and its competitors'. The panel does not retire that work. It surrounds it. It is the layer that lets a team ask a hundred more questions in the months between the studies that take months.

## Testing the decision before the budget

The other thread Kim keeps returning to is prediction. Most teams cannot afford to pre-test a television campaign the traditional way, so they don't — and they learn whether it worked after the money is gone. A persona panel reverses that order. You show it a new idea or a draft script and ask the questions that normally only arrive after launch: what is missing, what lands, what would you cut.

The same logic runs into the music — which is where MusicDatak comes in. Kim describes getting back the fifty tracks his listeners already love, and then the ten more they would love that the station is not playing yet. That last ten is the whole point. It is a call that used to rest entirely on a programmer's instinct, now made testable.

None of this is a pitch. It is a working description of how one radio group now decides: faster, across more questions, with a model of its own audience it can question at any hour. The traditional research still happens. There is simply more intelligence around it.

That is the standard I hold all of this to. Not whether the AI sounds impressive — whether it lets a good leader make the call sooner, and defend it afterwards. On the evidence of this conversation, NGroup is there.

— Samuel, Paris, May 2026