Social Identity Theory at Work — Why 'Us vs Them' Quietly Runs Your Org

A 200-person company I worked with had a brutal 'product vs engineering' war. Same office. Same equity. Same mission slides. The split formed within 18 months and nobody designed it that way — it just appeared, the way mold appears in a damp wall. By the time the leadership team noticed, it was structural: separate Slacks, separate offsites, separate stories about who was the problem.

Why it matters

Henri Tajfel's 1970s minimal-group experiments at Bristol showed something disturbing: even random, meaningless assignment to groups (literally coin-flip groups based on a fake preference for one painter over another) was enough to produce in-group favoritism and out-group discrimination. The implication for workplaces is huge — functional silos, tenure tribes, HQ/remote splits, founder/joiner splits, even desk-cluster identities aren't 'character flaws' or 'culture failures'. They are predictable physics.

You will get group identity formation in your org whether you design for it or not. The only question is whether the architecture you build channels it productively (think Pixar's directors, sports team rivalries) or toxically (Sales hates Product, HQ resents Remote, old-timers resent the new hires).

The theory

Tajfel and Turner's social identity theory says people move through three steps automatically: classify (we vs they), identify (I'm a we), and compare (we're better). The comparison generates self-esteem when the in-group looks good — so the brain finds ways to make the in-group look good, often at the out-group's expense. This isn't malice; it's the basic operating system.

Once a salient boundary exists — any visible difference that the group can rally around — the bias activates. Functional roles, tenure, location, founder status, and even Slack channel membership are all enough to trigger it. The boundary doesn't need to be meaningful; it just needs to be visible.

“We do not need to know much about another group to dislike them. We only need to know that they are 'them'.”

Three moves to manage it

The classic workplace splits

Example

Pixar's 'Braintrust' (see our case study) deliberately mixes directors across films so every project gets cross-pollinated feedback. The superordinate identity is loud and explicit — 'we make great films', not 'we're on Toy Story 4'. New directors are pulled into other directors' rough cuts and asked to critique honestly; senior directors do the same on the new director's work. The architecture beats the org chart.

Compare with companies where each product team is told 'you are a startup within a startup'. That language sounds empowering but Tajfel predicts exactly what happens: each team builds its own tools, its own rituals, its own mythology, and within two years the company is a federation of warring city-states held together by the org chart and very little else.

Apply on Monday

• Identify the two strongest sub-identities in your org. Are they pulling toward or away from each other?
• Audit your cross-team rituals — do they reinforce 'us' or just hand off work?
• Set one OKR this quarter that requires two functions to win together (and lose together).
• Watch your own language: count 'they' vs 'we' patterns in your last five meetings — your team copies your prepositions.
• Make at least one decision this quarter in a forum that mixes sub-groups deliberately.

Common mistakes

• Trying to dissolve sub-identities — people resist because identity is self.
• Assuming a culture deck solves it — architecture beats slogans every time.
• Hiring a 'culture lead' to fix what is actually a structural identity problem.
• Ignoring HQ/remote, tenure, and founder/joiner splits — all classic Tajfel triggers.
• Building offsites that reinforce the sub-group instead of the superordinate identity.
• Letting joint OKRs become 'lead team's OKR + a sidecar' — joint must mean shared accountability.

Reflection prompts

1. Which two groups in your org need a superordinate identity reinforced now?
2. Where are your rituals deepening silos accidentally?
3. What's a joint goal you could set this quarter that forces 'we'?
4. How often do you publicly celebrate cross-team wins vs single-team wins?

Takeaways

• Group identity forms automatically — design it on purpose.
• Sub-identities are not the enemy; missing superordinate identity is.
• Joint OKRs with shared accountability are the single highest-leverage architectural move.
• Your language is contagious — watch your 'we' vs 'they'.