The Glue Person

The no-stats all-star

In 2009, Michael Lewis wrote a profile for the New York Times Magazine about Shane Battier, a basketball player for the Houston Rockets. Battier’s individual statistics were thoroughly mediocre. He was not a great scorer, not a standout rebounder, not an exceptional athlete by NBA standards. In a league obsessed with individual performance metrics, he was unremarkable.

But every team Shane Battier played on won more games than expected. When he was on the court, his teammates shot better and his opponents shot worse. When he sat out, the effect reversed. The pattern was consistent across teams, seasons, and coaches. The data was unambiguous: Battier made every team he joined measurably better.

Lewis called him “The No-Stats All-Star.” Battier’s contribution was invisible to traditional box scores because it operated at the team level, not the individual level. He took charges. He contested shots without fouling. He positioned himself to help teammates recover. He communicated defensive assignments. He did the unglamorous, untracked work that shifted the probability of winning.

Daryl Morey, the Rockets’ analytics-driven general manager, valued Battier precisely because the team-level data revealed what the individual data concealed. The question Morey asked was not “how good is this player?” but “how much does this player improve the team?” These are fundamentally different questions — and the investment industry almost exclusively asks the wrong one.

What is a glue person?

A glue person is someone whose individual output is hard to measure but who elevates everyone around them. Their contribution shows up in team-level outcomes, not personal metrics. They are the connective tissue that turns a group of talented individuals into a functioning unit.

In an investment context, the glue person is the one who:

• •Asks the question in a meeting that reframes how everyone else thinks about a stock
• •Connects two analysts who are working on related ideas but don’t know it
• •Defuses tension between team members before it becomes a rift
• •Mentors junior team members in a way that accelerates their development by years
• •Builds and maintains the team’s analytical infrastructure — the databases, the tools, the shared frameworks
• •Creates a culture where people feel safe to challenge each other’s ideas
• •Remembers the precedent from three years ago that changes the framing of today’s decision
• •Translates between different thinking styles so that a quantitative insight reaches a qualitative thinker and vice versa

None of these contributions show up in a performance review that measures AUM, personal stock picks, or number of company meetings attended. The glue person’s value is legible only at the team level — in the quality of discussion, the speed of information flow, and the consistency of collective output.

Battier’s case is the most dramatic version of this because his individual stats were mediocre — the gap between his personal metrics and his team impact was so large that it became the story. But glue behaviour does not require weak individual performance. Some of the best glue people are also strong individual contributors. The point is not that glue people can’t be stars — it’s that their most valuable contribution is the one that never shows up in the individual numbers. A brilliant analyst who also connects the team, mentors juniors, and raises the quality of collective debate is worth far more than their stock picks alone would suggest. The tragedy is that most evaluation systems only see the stock picks.

Why traditional metrics miss them

The investment industry measures what is easy to measure: individual track records, assets under management, number of ideas generated, meetings attended, notes published. These metrics reward the visible, the individual, and the quantifiable. They systematically miss the invisible, the collective, and the qualitative.

This is the same error that basketball made before analytics. For decades, the sport rewarded scorers — the players whose contribution was most visible in the box score. Players like Battier, whose contribution was diffuse and team-level, were undervalued. It took a new generation of analysts, willing to look at team-level data and plus/minus statistics, to identify the players who actually shifted the win probability.

The investment industry is still in the pre-analytics era when it comes to valuing team contribution. Most firms reward the person who pitched the winning stock. Almost none systematically ask: “When this person joined, did the team’s collective output improve? When they left, did it decline?” Until firms start asking that question, glue people will continue to be undervalued, under-compensated, and eventually lost to environments that do not deserve them.

Designing for glue: what DeepMind understood

Google DeepMind took this idea further. Rather than leaving team composition to chance, they intentionally paired deep specialists — researchers who were technically brilliant but not natural connectors — with people whose primary skill was making teams work. The connectors did not need to be the strongest researchers. Their role was structural: to ensure information flowed between subgroups, to translate between different technical languages, and to maintain the cohesion that allowed the specialists to do their best work.

The insight is that glue is not an accident of personality — it is a design choice. You can build teams that are structurally connected by deliberately including people whose contribution is connective rather than individually productive. The alternative is a team of brilliant siloes that never quite adds up to more than the sum of its parts.

The interactive network below illustrates this. A team of specialists naturally clusters into subgroups. Add a glue person and watch the information flow transform. Increase the number of connectors and the network becomes resilient — if one is absent, the others maintain the bridges.

How to identify and protect them

Recognising glue people requires looking at different signals:

• •Team-level outcomes, not individual metrics. When did the team improve? When did it decline? Who was added or removed at those inflection points?
• •Quality of discussion. Are meetings more productive when this person is in the room? Do ideas get sharper, debates get more honest, decisions get clearer?
• •Information flow. Does this person connect people who would otherwise work in siloes? Do they translate between different thinking styles or specialisms?
• •Cultural contribution. Does this person raise the standard of behaviour? Do they model the kind of challenge, honesty, and intellectual generosity that the team aspires to?
• •Counterfactual thinking. If this person left tomorrow, what would break that is not captured in any dashboard?

Once identified, glue people must be valued explicitly. This means compensating them for their team-level contribution, not just their individual output. It means protecting them from a culture that rewards only the visible. And it means recognising that losing a glue person is not like losing any other team member — it is a structural loss that degrades the performance of everyone who remains.

The asymmetry

There is a painful asymmetry in how glue people are treated. The star analyst who generates one exceptional idea gets the bonus, the promotion, and the profile. The glue person who created the environment in which that idea could surface — who built the database the analyst queried, who asked the question that reframed the thesis, who resolved the team conflict that was blocking collaboration — gets nothing. Or worse, gets told their contribution is hard to measure and therefore hard to reward.

This asymmetry is self-defeating. The firms that recognise and reward glue behaviour will build teams that compound their collective intelligence over time. The firms that do not will keep hiring stars, keep losing glue people, and keep wondering why their teams never quite gel.