Culture is not the values statement on your wall. It is the set of behaviours your organisation actually rewards, tolerates and punishes — visible in who gets promoted, which shortcuts go unremarked, and what people do when no leader is watching. If your stated value is "customer first" but the fastest route to promotion is hitting a shipping number regardless of returns, your real culture is "ship it and move on." Everyone already knows this. A transformation only works when you change the systems that produce the behaviour, not the words that describe the aspiration.

That makes culture a COO problem, not just an HR one. You own the operating cadence, the performance metrics, the budget, and most of the cross-functional wiring. Those are the levers that actually move behaviour. A poster campaign changes nothing; changing what gets measured, funded and promoted changes everything.

The payoff is worth the effort. Organisations that genuinely align stated values with rewarded behaviour tend to see lower regretted attrition, faster decisions, and fewer of the quiet workarounds that erode operational quality. Below is a sequence you can run — assess honestly, define behaviours (not adjectives), change the systems, handle resistance as information, and measure the leading indicators before you claim victory. Culture is only one pillar; structure, talent, and culture together make up the wider picture set out in the organizational excellence guide.

Assess the culture you actually have, not the one you wish for

Most culture assessments fail because they ask people to rate abstractions. "Do we value integrity?" gets you a meaningless 4.2 out of 5. Instead, look at behaviour and at the gap between what's said and what's done.

Strong assessment triangulates three sources. Observed behaviour: sit in on decision meetings and note who speaks, who gets overruled, and how disagreement is handled. System evidence: pull the last two years of promotions and ask what those people actually did to earn them — that is your real values statement. Direct voice: run short, anonymous pulse surveys plus a handful of skip-level and exit interviews, and read exit data honestly rather than filing it.

Weak assessment looks like a single annual engagement survey that leadership skims and shelves. Strong assessment produces a specific, uncomfortable list: "We say we empower managers, but every purchase over $500 needs three approvals" or "We reward the loudest voice in the room, so quiet analysts stop contributing." When you can name the behaviour, the mechanism that rewards it, and the gap it creates, you have something you can change. Vague findings like "communication could be better" are a sign you stopped one layer too shallow. If your data is thin, an honest operations assessment framework can force the specificity.

Define behaviours, not adjectives

"Be collaborative" is unmeasurable and everyone thinks they already are. The work of transformation is translating each aspiration into two or three observable behaviours that a new hire could copy and a manager could coach against.

Take "customer-obsessed." As a behaviour standard it becomes: every product decision cites a specific customer problem; support tickets over 48 hours old are escalated automatically; engineers spend one day a quarter on support calls. Those are things you can see, count, and hold people to. "Customer-obsessed" as an adjective is a mood; as a set of behaviours it is a system.

The right-hand columns are what you build the transformation around. This is also where a COO's operational instinct beats an HR-led effort: you are used to turning vague intent into a repeatable process, and culture change is exactly that.

Change the systems, because systems beat slogans

Behaviour follows incentives. If you want the new behaviour to stick, you must change the four systems that quietly teach people how to survive and get ahead: who you hire, who you promote, what you measure, and how you meet.

Hiring. Add the target behaviours to your interview scorecard and reject on them even when the skills are strong. A brilliant engineer who dismisses colleagues in the interview will do the same on the team, and every existing employee will read the hire as permission. Promotion. This is the loudest signal you send. Promote one person who embodies the new behaviour and explain publicly why; refuse to promote a top performer who tramples it, and explain that privately. People believe promotions, not memos. Metrics. Retire the single number that drives the old behaviour, or balance it — pair a shipping-speed metric with a quality or returns metric so speed can't be bought with defects. Meetings. Your operating rhythm is culture in miniature. If your weekly review only ever asks "did we hit the number," you are training people to hide problems until they explode.

Strong looks like a COO who redesigns the quarterly review so that a blameless "what did we learn from the miss" section carries equal weight to the results section. Weak looks like a values launch event followed by an unchanged promotion process — the fastest way to teach cynicism. Aligning change management strategies with these system levers is the difference between a transformation and a memo. The tooling to support it — trackers, dashboards, feedback channels — is real but secondary; pick your change management tools after you've decided which behaviours the systems must reward, not before.

Handle resistance as information, not defiance

Resistance is data. When a team pushes back on a new standard, the useful question is not "how do I overcome them" but "what do they know that I don't?" Sometimes the answer is that the new behaviour conflicts with a metric you forgot to change — the resistance is rational and the fix is upstream, in the system.

Weak change leaders treat resistance as an attitude problem to be crushed with more communication. Strong ones separate three types. Rational objections ("this new approval flow adds two days to every order") point to a design flaw — fix it and you gain an ally. Loss-based resistance (someone's status, expertise or comfort is threatened) needs honesty about the trade and, where possible, a role in the new world. Capability gaps (people want to comply but lack the skill) need training, not pressure; this is where employee engagement and real talent development do the heavy lifting.

A concrete example: a COO rolling out "decisions at the lowest sensible level" finds middle managers stalling. Digging in, she learns they've been burned before — earlier autonomy came with blame when things went wrong. The real fix isn't a stern email about empowerment; it's a public commitment that reversible decisions won't be second-guessed, backed by not second-guessing the next one that goes sideways. Behaviour changes when the risk of the new behaviour genuinely drops.

Sequence the work in phases and don't declare victory early

Culture change is measured in quarters and years, not weeks. A useful frame is four overlapping phases, each with a different job. Rushing from launch to "done" is the classic failure — the behaviour looks changed while a leader is watching, then snaps back the moment attention moves on.

Two disciplines keep phases honest. First, recruit culture champions — respected peers, not just managers — who model the behaviour and coach it laterally; peer influence moves a culture faster than top-down mandate. Second, run the change through your existing operating cadence rather than as a side project. Threading it through cross-functional teams and your normal reviews means it competes for real attention instead of dying as an initiative nobody has time for.

Measure leading indicators, not just the annual score

The trap is waiting a year for the engagement survey to tell you whether it worked. By then you've lost the ability to course-correct. Track a small set of leading indicators monthly or quarterly so you can see behaviour shifting — or not — while there's still time to adjust.

Useful leading signals include: the share of promotions that visibly reward the new behaviour; regretted attrition among your strongest performers (rising means the old culture is still winning); decision cycle time on reversible calls; the proportion of post-mortems that are genuinely blameless; and pulse-check responses to two or three behaviour-specific questions repeated often. Lagging indicators — retention, customer satisfaction, financial performance — matter, but they confirm the change months after the fact. Tie the whole effort to your existing COO success metrics so culture isn't a separate scoreboard nobody looks at, and make the link to operational excellence explicit: the point of changing behaviour is better, faster, more reliable work, and that is what you're ultimately measuring.

Finally, embed the change where it can't quietly reverse. Update performance review criteria to include the behaviours, write them into onboarding, and factor cultural fit into succession planning so the next generation of leaders is chosen partly for how they work, not only what they deliver. A culture only becomes permanent when the systems that select and reward people would produce it again even if you left.

Key takeaways

• Culture is behaviour that gets rewarded, not values on a wall. Change what you hire for, promote for, measure and meet about — not the poster.
• Translate every value into two or three observable behaviours. "Accountable" is unmeasurable; "decisions have a named owner and a review date" is coachable and countable.
• The four systems that teach culture are hiring, promotion, metrics and meetings. Promotions are the loudest signal you send; people believe them over memos.
• Treat resistance as information. It usually points to a system you forgot to change, a real loss to be handled honestly, or a skills gap to train — not defiance to crush.
• Sequence in phases and don't declare victory early. The change is real only when the new behaviour survives a stressful quarter.
• Track leading indicators monthly — promotion patterns, regretted attrition, decision speed, blameless post-mortems — rather than waiting a year for the engagement survey.

Frequently asked questions

How long does a cultural transformation realistically take? Early behaviour shifts can appear within a few months once you change a visible system such as a promotion or a metric. Genuine embedding — where the new behaviour survives a bad quarter without leaders policing it — usually takes one to three years. Anyone promising a culture "fixed" in a quarter is describing a launch event, not a change. Should the COO or HR lead this? Both, in different roles. HR brings the assessment tools, training and change expertise; the COO owns the operational systems — metrics, budget, operating cadence, cross-functional wiring — where behaviour is actually rewarded. Because those systems are where culture lives, a COO-led effort partnered with HR tends to stick better than an HR-only initiative that can't change what gets measured or promoted. What is the single biggest reason culture change fails? A mismatch between the stated change and the promotion and metric systems. If you announce new values on Monday and promote someone who violates them on Friday, you've taught the whole organisation that the values are decoration. The reverse — one well-explained promotion that rewards the new behaviour — moves culture more than any communication campaign. How do we handle a top performer who won't adopt the new behaviour? This is the defining test. If you exempt your best salesperson or engineer from the standards, everyone learns the standards are optional for the valuable. Coach directly and specifically first; if the behaviour doesn't change, be willing to act, because keeping them on your terms costs you the credibility of the entire effort. Handle the loss of talent honestly, but don't let one person hold the culture hostage. How do we measure something as soft as culture? Measure behaviour and systems, which are concrete. Count the share of promotions that reward the target behaviour, track regretted attrition and decision cycle time, and repeat two or three behaviour-specific pulse questions often. These leading indicators move well before lagging outcomes like retention and customer satisfaction, so you can course-correct instead of waiting a year to find out it didn't work. Do we need to change everyone at once? No — and trying to usually backfires. Start with the systems and a group of respected culture champions who model the behaviour and influence peers laterally. Behaviour spreads through credible colleagues faster than through mandates, so a strong early cohort plus aligned promotion and metric systems will pull the middle of the organisation along without a mass mandate that mostly generates cynicism.