What Hockey Can Teach Us About Customer Segmentation: 2026 Edition

The start of the Canadiens’ season always brings a certain buzz to the Numea office. Last year, we built a segmentation of NHL players, which means grouping players into typical profiles based on their stats, just like we do for our clients. We did it for fun first, but also to show what a segmentation can actually be used for. Read the first edition here!

Given the interest the project got last year, we decided to rework the segmentation model more seriously. We asked our in-house “experts” for their input on choosing and defining the variables, and our model is now even more robust than last year’s.

First, the data

The key to a good segmentation, and to any AI project for that matter, is data. First, you need access to clean, centralized data. In this case, we used 2025-2026 stats for 828 NHL players (source: MoneyPuck.com).

Next, as anyone who has worked on a segmentation project with us knows, the first step is to choose the key concepts that will group similar players together.

In our first edition, the variables were used as is. For example, we looked at a player’s total goals for the season. This year, we measured performance while the player is on the ice (per minute played), while still taking his total ice time into account. That way, a fourth-liner isn’t at a disadvantage against a star who plays 22 minutes a game.

We also added new concepts to better tell players apart. These indicators cover both offensive production and defensive playing style.

Here are the model’s variables, in order of importance:

The segments

Once the variables were chosen, we grouped together the players who are most alike. The result: six segments, split into three groups: forwards, defensemen and fill-in players. Each segment plays a different role on the team, just like each type of customer in your business.

Forwards

Defensemen

Fill-in players

The Demidov case

Last year, Ivan Demidov was our example of a young Canadiens rookie. This year, with 62 points, he’s one of the offensive stars. Just like a customer, a player can move from one segment to another. That’s why it matters to track these movements over time!

Ice time, points and segment size

The chart below places each segment on two criteria: average ice time per game and points per game. Bubble size represents the number of players in each segment (the bigger the bubble, the larger the segment). Offensive stars clearly stand out at the top, while offensive defensemen stand out for their ice time.

Note: depth forwards and fill-in players look alike on these two criteria. What really sets them apart is the number of games played.

A portrait of the league

Not all segments carry the same weight. Offensive stars make up only 14% of players, yet they collect 35% of the league’s points. Add supporting forwards, and 42% of players account for 68% of points.

Your customers work the same way: a minority often generates most of your revenue. You just need to know who they are!

Every team, a different strategy

Looking at each team’s makeup, you can see very different strategies.

• Colorado relies on its offensive stars, who make up 20% of its roster. Columbus, on the other hand, has none and leans on depth instead, with 42% supporting forwards.
• Dallas can count on a solid defensive corps, with 31% stay-at-home defensemen, compared to 17% league-wide. Calgary has almost as many (29%), but no offensive stars and very few offensive defensemen (4%), which makes for a team that struggles to score.
• The two Stanley Cup finalists (Carolina and Vegas) share an interesting trait: plenty of offensive defensemen (20% and 25%, vs. 13% league-wide) and few stay-at-home defensemen (8%, vs. 17%).
• And the Canadiens? Their lineup is very close to the league average, with four offensive stars: Suzuki, Caufield, Slafkovsky and Demidov. That balance is a change from last year, when the team stood out mostly for its young players in development.

Marketing works the same way: some companies bet on their most profitable customers, while others build on a broad, stable base. What matters is knowing which mix you have, and above all, which one you want!

What about your customers?

Replace “players” with “customers,” and the findings still hold.

• A minority accounts for the majority. Offensive stars make up 14% of players, but 35% of points. In most businesses, a small share of customers generates most of the revenue. Do you know who they are?
• Segments shift. Demidov went from rookie to star in a single year. Your customers’ behaviour changes too. A high-potential customer today can become your best customer tomorrow, or drift away if no one pays attention.
• Every segment has a role. Fill-in players barely score, but they round out the lineup. Your occasional customers don’t bring in as much as your best ones, but they don’t deserve to be ignored. They deserve an approach that fits.
• There’s no single winning formula. Colorado bets on offence, Dallas on defence, Vegas and Carolina on offensive defensemen. What matters is knowing the makeup of your customer base and making choices accordingly.

In practical terms, segmentation shows you where to focus your efforts: retaining your best customers, helping those with potential grow, and adapting your approach for everyone else. A good GM knows the role of every player on the roster. Do you know the role of every one of your customers?

At Numea

We segmented NHL players for fun, but the method behind it is the same one we use every day with our clients’ data.

We may not be able to predict who’ll win the Stanley Cup this year, but we can definitely help you get to know your customers better.

Discover our segmentation services

Want to segment your customers? We can help you turn data into action.

Contact us