Ask a leader to describe "the culture" and they'll give you a single answer, usually a flattering one. Then walk the floor. The engineers ship on their own cadence and quietly ignore the all-hands; the salespeople live by the quarter; the customer-support team has a warmth the rest of the building never sees. That isn't a culture failing to take hold, it's the normal state of any organisation past a certain size. And it's why merging two companies is so much harder than the deal model admits: you aren't blending two cultures, you're trying to settle the fights between many.

The quick version

  • One culture is a myth. Every organisation is a federation of subcultures, by function, location, profession and tenure. The dominant culture is just the loudest, not the only.
  • Subcultures come in three flavours. Some reinforce the dominant values, some hold their own without conflicting, and some actively push back. Only the third is a problem, and even then, not always.
  • There's no single "right" way to combine cultures. Two firms can integrate, one can absorb the other, they can stay separate, or it can all dissolve. The wrong choice, or two firms wanting different choices, is what creates the friction people call a "culture clash."
  • Culture is the quiet deal-killer. In Bain's 2023 survey, nearly half of acquirers named cultural fit or management-team integration as a reason past deals fell short, and most still struggle with it despite knowing it matters.

The idea in depth

Your organisation is plural, not singular

The single most useful correction to most culture talk is grammatical: stop saying "our culture" and start saying "our cultures." In a much-cited 1983 analysis, Joanne Martin and Caren Siehl argued that the picture of a monolithic corporate culture handed down from the top is a convenient fiction. Real organisations grow subcultures, pockets of shared meaning that form around a function, a site, a profession, a founding era, and they can coexist with the dominant culture in very different ways.

Martin and Siehl sorted them into three types. An enhancing subculture holds the dominant values even more fervently than the average employee (think of a flagship team that embodies "the way we do things" to the point of evangelism). An orthogonal subculture embraces the dominant values and adds its own distinct ones that don't conflict, the finance team that's fully bought into the mission while also keeping its own exacting professional norms. A counterculture holds values that directly clash with the core; their famous example was John DeLorean's maverick unit inside the buttoned-down General Motors of the era.

So the move is: before you try to "change the culture," map the subcultures you actually have. Name the three or four that matter, and for each, ask which type it is. You manage an enhancing subculture by feeding it, an orthogonal one by leaving its distinct norms alone, and a counterculture by deciding, deliberately, whether it's a problem to fix or a useful source of dissent to protect.

The limitation: these categories are a lens, not a microscope. The same subculture can look enhancing on the values everyone agrees on and countercultural on the one they don't, and a unit's type can flip when leadership or strategy changes. Use the labels to start a conversation, not to file people permanently.

This is also why psychological safety is so uneven inside one company. Amy Edmondson's foundational 1999 study established that the willingness to speak up without fear is a team-level climate, not a company-wide trait, it can be high in one group and low in the next, set largely by the local manager. Safety lives in the subculture. A merger that ignores this and announces "we have a culture of openness" from head office is making a promise the org chart can't keep.

flowchart TD
    A(["The dominant culture
(the official story)"]) --> B(["Enhancing
holds the values
even harder"])
    A --> C(["Orthogonal
shares the values +
its own, no conflict"])
    A --> D(["Counterculture
values that
directly clash"])
Three ways a subculture relates to the dominant culture. After Martin & Siehl (1983). Leaders Loop

A merger doesn't merge cultures, it forces an acculturation choice

Here's the part deal teams skip. When two organisations come together, there is no automatic "blend." Someone has to decide how the cultures will relate, and the cleanest model for that comes from Afsaneh Nahavandi and Ali Malekzadeh, whose 1988 paper in the Academy of Management Review applied the psychology of acculturation to mergers. They describe four possible modes. In integration, both firms change and a genuinely new shared culture emerges, with each side keeping much of its identity. In assimilation, the acquired firm gives up its culture and adopts the acquirer's, common, and workable when people see the acquirer's way as better. In separation, the acquired firm is left largely alone, with minimal cultural exchange. And in deculturation, the acquired culture is dismantled but nothing coherent replaces it, the worst outcome, leaving people adrift.

Their central insight is about expectations. Trouble comes not just from the mode chosen, but from a mismatch: when the acquired employees want, say, separation (keep our identity) and the acquirer is quietly planning assimilation (you'll do things our way now). That gap is what produces acculturative stress, the resentment, disengagement and quiet exits that read, from the outside, as a "culture clash."

So the move is: pick the mode out loud, early, and check whether both sides expect the same one. If the thesis is "we bought them for their distinctive way of working," you've chosen separation or integration, and you must say so and protect it, because drifting by default into assimilation destroys the asset. If the thesis is "we're bringing them onto our platform and our standards," that's assimilation, say that, honestly, so people aren't promised an autonomy they won't get.

quadrantChart
    title How will the two cultures relate?
    x-axis "Acquired firm keeps little" --> "Acquired firm keeps much"
    y-axis "Low contact between firms" --> "High contact between firms"
    quadrant-1 "Integration: both change, new shared culture"
    quadrant-2 "Separation: leave them alone, low exchange"
    quadrant-3 "Deculturation: old culture gone, nothing replaces it"
    quadrant-4 "Assimilation: they adopt the acquirer's culture"
The four acculturation modes. The damage comes from the two firms expecting different ones. After Nahavandi & Malekzadeh (1988). Leaders Loop

Why culture is the quiet failure mode, and what beats it

The evidence that culture decides deals is consistent, and a bit uncomfortable. EY frames culture as a lever you have to manage on purpose to avoid the widely-cited 70–90% M&A failure range, value created when it's understood, destroyed when it's ignored. Bain's 2023 work on cultural integration is more pointed. Nearly half of the acquirers it surveyed named cultural fit or management-team integration as a reason past deals had fallen short. And although culture was an early focus for around 80% of integrations, roughly 75% of acquirers still hit cultural problems serious enough to need real intervention. Knowing it matters, it turns out, isn't the same as handling it.

Bain's practical recommendation is the one worth stealing: don't try to align two whole cultures, which can take years and may never happen. Instead, find the specific cultural fault lines, the few concrete differences in how the firms make decisions, run meetings, judge performance and treat people, and resolve those. It's the same instinct that makes the subculture map useful: you don't integrate "the culture," you integrate the handful of everyday practices where the two sides genuinely operate differently.

The practical step: in the first weeks, run a short, honest comparison of how each firm actually does five or six ordinary things, how a budget gets approved, how disagreement gets escalated, how someone gets promoted, how fast email gets answered. For each, decide explicitly whose way wins, or whether both persist. The ambiguity in those small rules is what your best people quit over, long before any grand "values" exercise lands.

You don't integrate "the culture." You integrate the dozen small rules where two companies disagree about how a Tuesday works.

The limitation: the headline failure statistics deserve a raised eyebrow. "Failure" is defined differently across studies, shareholder return in some, stated synergies in others, and culture is hard to isolate from price paid, market timing and plain execution. Read the numbers as a strong directional warning that culture is load-bearing, not as a precise law.

A worked example

The following figures are illustrative, to show the mechanics, not a real company.

Picture a 600-person enterprise software firm, Meridian, acquiring a 60-person design-led product studio, Kite, for £35m. The deal thesis is explicit: Meridian is buying Kite's craft culture and its senior designers, because Meridian's own product has a reputation for being powerful and unloved. On the Nahavandi–Malekzadeh map, that thesis points squarely at separation early, drifting toward integration later, keep Kite intact, then slowly let the two influence each other.

The integration lead does three things. First, she names the mode out loud, to both sides, so Kite's founders hear "we are not absorbing you" from someone with authority, and Meridian's engineering leaders hear "no, you do not get to put them on your sprint process in month one." Second, she maps the subcultures rather than treating each firm as one block. Kite turns out to be largely an orthogonal subculture-in-waiting: fully able to share Meridian's mission while keeping its distinct design norms. But Meridian's own platform-engineering group is a quiet counterculture that prizes stability and views Kite's "ship and learn" instinct as recklessness, so that seam, not the headline merger, is the real fault line. Third, she resolves the concrete fault lines: within a month the two sides agree that Kite keeps its own release cadence and design reviews, but adopts Meridian's security and on-call standards, with the trade documented so no designer is guessing on day 20.

The avoided disaster is deculturation: the version where Meridian, meaning well, dissolves the Kite brand, scatters the designers across existing teams, and imposes its process, and within a year the senior designers it paid £35m for have all left, taking the only thing the deal was for. Naming the mode, and protecting the right subculture, is what keeps the asset alive.

Frequently asked questions

Aren't subcultures just a sign of a weak or divided culture?

No, they're a sign of a normal one. Any organisation big enough to have departments will grow subcultures, because the work itself differs: a legal team and a growth team face different problems and reward different behaviours. The goal isn't to stamp them out (you can't, and the orthogonal ones are often where real expertise lives) but to know which is which. A counterculture worth worrying about is one whose values block the strategy; a counterculture worth keeping is one that supplies the dissent that stops groupthink.

What's the difference between this and post-merger integration generally?

Post-merger integration is the whole job of making two companies one, systems, brand, org structure, finance, the lot. Cultural integration is the human layer underneath it, and it's the part that most often quietly fails. You can migrate every system on schedule and still lose the deal if the acquired team's best people leave because nobody decided, out loud, how the two cultures would relate. See Post-merger integration for the broader sequence.

Should we always aim for "integration", a brand-new blended culture?

Not necessarily, and assuming so is a common mistake. Integration sounds the most enlightened, but it's the hardest and slowest, and it's the wrong choice when you bought a company precisely for a distinct culture you should protect (that's a case for separation). Assimilation is honest and fast when one firm's way is genuinely stronger and people accept it. The failure mode is deculturation, and you mostly back into that by choosing nothing and letting the bigger firm's habits quietly win.

How do you actually measure a culture before a deal?

Imperfectly, which is why "cultural due diligence" is honest about being directional rather than precise. You're not scoring a single number; you're looking for the fault lines, concrete differences in how decisions get made, how performance is judged, how fast things move, how people talk to each other. Interviews, observing real meetings, and a few targeted survey questions beat a generic "values" deck. The aim is to find the half-dozen places the two firms genuinely differ, before they become resignations.

Who owns cultural integration?

It needs a named owner with authority and air cover, not a committee, and not "HR will handle the soft stuff." Culture decisions are operating decisions (whose approval process wins, whose promotion bar applies), so they need someone who can actually make calls, working alongside the integration office. Treating culture as a workstream with a deadline, rather than an offsite, is the difference between handling it and hoping.

Related in the Toolkit

Where to go next