Values statements are not what moves performance. The evidence puts most of the variance somewhere far less glamorous, and it is the layer most culture programmes reach last.
Contents
- The short answer
- Key takeaways
- What is a high-performance culture?
- Does it actually pay?
- Why the standard roadmap under-delivers
- The layer where it is won or lost
- The 2026 complication
- A five-stage roadmap
- What to measure, and when
- Culture problem or management-system problem?
- Where MASSIVUE fits
- Frequently asked questions
- Related reading
- Sources
The short answer
You build a high-performance culture by fixing the manager layer first, then changing the systems that decide what gets rewarded, and only then codifying any of it as values. Gallup estimates that managers account for at least 70% of the variance in employee engagement between business units, which means the difference between your best and worst teams is mostly a management difference, not a company-wide culture difference. Most roadmaps invert that order. They start with an executive statement of values and reach the manager layer last, if at all.
Key takeaways
- The variance is inside your company, not between you and your competitors. Gallup's meta-analysis covers 183,806 business units precisely because the unit of analysis that predicts performance is the team, not the enterprise.
- The performance link is real and quantified. Top-quartile teams beat bottom-quartile teams by a median 23% on profitability and 78% on absenteeism, across 347 organisations in 90 countries.
- The manager layer is currently deteriorating. Global manager engagement fell nine points between 2022 and 2025, and five of those points came in a single year.
- It is a solvable problem, not a macro one. In organisations Gallup classes as best practice, 79% of managers are engaged, against a global average of 22%.
- The cheapest available intervention is the one most organisations have skipped. Only 44% of the world's managers say they have received any management training.
- AI has changed the work without changing the management system. Only 12% of employees strongly agree AI has transformed how work gets done, and 89% of executives report no measurable productivity effect over the past three years.
- Values come last, not first. Codify the behaviour you have already produced. Announcing behaviour you have not produced yet teaches people that the statement and the reality are unrelated.
What is a high-performance culture?
A high-performance culture is a pattern of behaviour, repeated across teams, in which people consistently do the work that produces results without needing to be told. It is observable in what happens under pressure, when there is no instruction and no supervisor present. It is not the values poster, and it is not the engagement score.
That definition has a practical consequence that most culture programmes ignore: if culture is a pattern of behaviour, it can only be assessed by observing behaviour and its consequences. Asking people to describe their culture measures their opinion of it, which is a softer and more manipulable quantity.
Three terms get used interchangeably here and should not be.
| Term | What it measures | How fast it moves |
|---|---|---|
| Culture | Repeated patterns of behaviour and interaction, including behaviour under pressure | Years, and only in response to changed conditions |
| Climate | How the environment feels right now, which shifts with events and leadership changes | Weeks to months |
| Engagement | Discretionary effort and intent to stay, measured at team level | Quarters, and it is the most reliable leading indicator you can get |
Engagement is the practical proxy, because it is the one with a published evidence base linking it to business outcomes at the level where you can actually act. That is what the next section is about.
Does it actually pay?
Yes, and the size of the effect is documented rather than asserted. This matters, because the culture field is unusually full of numbers with no traceable source.
Gallup's Q12 meta-analysis, eleventh edition, pooled 736 research studies across 347 organisations in 53 industries and 90 countries, covering 183,806 business and work units and 3,354,784 employees. It used Hunter-Schmidt methods to correct for sampling and measurement error, and reported a true score correlation of 0.49 between engagement and composite performance. Business units in the top half on engagement more than double their odds of success against those in the bottom half.
The median percentage differences between top-quartile and bottom-quartile units are the practical headline.
| Outcome | Median difference, top quartile vs bottom quartile |
|---|---|
| Absenteeism | 78% |
| Wellbeing (thriving employees) | 70% |
| Safety incidents | 63% |
| Turnover, low-turnover organisations | 51% |
| Quality (defects) | 32% |
| Shrinkage (theft) | 28% |
| Profitability | 23% |
| Turnover, high-turnover organisations | 21% |
| Productivity (sales) | 18% |
| Productivity (production records) | 14% |
| Customer loyalty and engagement | 10% |
Why the standard roadmap under-delivers
The conventional sequence is familiar because almost everyone publishes a version of it, including earlier versions of this article. Assess the current culture. Define a vision and a set of values. Build strategies and initiatives. Implement. Measure and improve.
Nothing in that list is wrong. The problem is the order, and what the order implies about where the work happens.
Starting with values puts the first and most visible effort into a company-wide artefact, at a level where the measured variance is smallest. It also front-loads the part of the programme with the weakest evidence behind it. Research by Donald Sull, Stefano Turconi and Charles Sull compared the published values of 562 companies against 1.2 million employee reviews and found no correlation between the values a company promotes and how employees rate it on those same values. Four of the nine values studied correlated negatively.
The second failure is subtler. A values-first sequence produces an early, highly visible commitment that the organisation has not yet earned. Every subsequent gap between the statement and daily experience is then read by employees as evidence that leadership either does not know what is happening or does not mind. That is an expensive way to start.
The layer where it is won or lost
Gallup estimates that managers account for at least 70% of the variance in employee engagement scores across business units. That single estimate reorganises the whole problem. If most of the spread between a strong team and a weak one is attributable to the person running it, then a culture programme that does not change what managers do is a communications exercise with a budget.
The uncomfortable part is what has happened to that layer recently.
Manager engagement, global
Gallup World Poll, 2025 fieldwork, 141,444 employed respondents across 140+ countries. The global average is the problem. The best-practice figure is the proof that it is fixable.
Global employee engagement itself fell to 20% in 2025, its lowest level since 2020, down from a 23% peak in 2022. Gallup puts the cost of low engagement at roughly $10 trillion in lost productivity, or 9% of global GDP. It attributes the recent decline primarily to managers, whose engagement fell while individual contributor engagement stayed broadly flat.
What actually moves manager performance
This is where the evidence gets unusually actionable, because the intervention is cheap and most organisations have not done it.
Fewer than half of the world's managers, 44%, say they have received any management training. Among those who have, Gallup finds half as many are actively disengaged as among the untrained. In a Gallup study of a manager training course focused on management best practices, participants showed up to 22% higher engagement than non-participants, the teams they led saw engagement rise by up to 18%, and manager performance metrics improved by 20% to 28%. Those results were measured nine to 18 months after training, which is long enough to rule out a short-lived enthusiasm effect.
Development also has to continue after the course. Training alone lifts manager thriving from 28% to 34%. Training combined with someone at work who actively encourages that manager's development lifts it to 50%.
The selection problem sits underneath the training problem
Gallup's longer-running research on manager selection found that companies fail to choose the candidate with the right talent for the manager role about 82% of the time, typically promoting on prior individual performance or tenure rather than on any assessment of managerial aptitude. Treat that figure as directional rather than precise; it rests on Gallup's proprietary talent assessment and has been carried forward across several editions of their work. The underlying point survives the caveat. If you promote your best individual contributors into management by default and then give 56% of them no training, the variance you observe between teams is not a mystery.
The 2026 complication
There is a reason a culture roadmap written in 2023 does not transfer cleanly to 2026, and it is not that the principles changed. It is that the work changed underneath them.
A survey of nearly 6,000 senior executives in the United States, United Kingdom, Germany and Australia, fielded between November 2025 and January 2026 by research teams at the Federal Reserve Bank of Atlanta, the Bank of England, the Deutsche Bundesbank and Macquarie University, found that 69% of businesses are actively using AI. In the same survey, 89% of executives reported no impact of AI on their own firm's labour productivity over the past three years, and more than 90% reported no impact on employment.
Gallup's data shows the same disconnect from the other side. Among workers in organisations that have implemented AI, 65% say it has had a positive effect on their individual productivity. Only 12% strongly agree that AI has transformed how work gets done in their organisation.
Individual gains that do not aggregate into organisational gains is a management-system signature, not a technology signature. The tools reached individuals. The processes, targets, hand-offs and review cycles that convert individual output into firm output did not change to match.
Two Gallup findings point at where that conversion is blocked. Frequent AI use is far higher among employees who strongly agree that AI integrates with their existing work systems, 86% against 52%. It is also far higher where managers actively support their team's use of it, 79% against 46%. Both levers sit below the executive layer and above the individual one. Both are manager and process questions.
A five-stage roadmap
This sequence is MASSIVUE's synthesis of the evidence above, applied through our enterprise transformation practice. It is deliberately the inverse of the conventional order on the two points that matter: it puts the manager layer near the front and values at the end.
Each stage below names one owner and one observable. If you cannot point at the observable, the stage is not complete, whatever the programme status report says.
| # | Stage | The question it answers | Owner | Observable that proves it happened |
|---|---|---|---|---|
| 1 | Baseline the variance, not the average | Which of our teams already perform the way we say we want to, and which do not? | CHRO with the CEO | Team-level engagement and outcome data with the spread reported, plus a stated minimum cell size |
| 2 | Fix the manager layer | Do the people who set the daily conditions know how to do it, and is anyone developing them? | CHRO, with each manager's own manager | Every people manager has had role training, and every one of them can name the person responsible for their development |
| 3 | Change what the system rewards | When doing the right thing costs someone their number, what happens? | CEO with CFO and CHRO | A named person was visibly rewarded for a decision that served the outcome and missed a target |
| 4 | Redesign the work, not just the tools | Has any process actually got shorter, or did we automate the same steps? | COO and process owners, with line managers | At least one end-to-end process with steps removed rather than accelerated, with the before and after cycle time recorded |
| 5 | Codify what you have produced | What is now reliably true about how we work here? | CEO | Every stated value has a specific behaviour from the last quarter that someone can point to |
Why stage 1 is about variance
An enterprise-wide engagement average is close to useless as a starting point. It tells you where you sit against a benchmark, which is a board slide, not a plan. The spread between your teams tells you something far more valuable: that some managers in your organisation, under your systems, with your constraints, have already solved the problem.
Those teams are the intervention. Studying them is cheaper, faster and more credible internally than importing a framework, because nobody can argue that it would not work here.
Why stage 3 cannot be skipped
Stages 1 and 2 will produce a visible improvement on their own, and many programmes stop there. The improvement then decays, because the reward system is still teaching the opposite lesson. People read incentives more accurately than they read messaging, and they will keep doing so.
Stage 3 is the hardest stage and the one with the shortest description, which is not a coincidence. It usually requires a single, specific, public decision rather than a policy: one instance where somebody was demonstrably better off for having done the right thing at a cost. Until that instance exists, everything above it is provisional.
What to measure, and when
Culture programmes tend to be measured with the instrument that is easiest to run rather than the one that would falsify the plan. This is a workable minimum set.
| Horizon | Measure | Read it as |
|---|---|---|
| Baseline | Team-level engagement spread, top decile against bottom decile | The size of the prize, and which managers to study |
| Baseline | Share of people managers with role training, and share with a named developer | Your two cheapest levers, quantified |
| Quarter 1 to 2 | Manager engagement and thriving, tracked separately from the workforce figure | Whether stage 2 is landing. This moves before team engagement does |
| Quarter 2 to 4 | Movement of bottom-quartile teams, not the enterprise average | The average can improve through attrition. Quartile movement cannot be faked that way |
| Quarter 2 to 4 | Cycle time on one redesigned process, before and after | Whether stage 4 removed work or just accelerated it |
| Quarter 3 onward | Voluntary regretted attrition in top-quartile teams | Your early warning that the reward system is still misaligned |
| Year 1 onward | Business outcome on the dimension you chose at baseline | The only measure that settles the argument |
Two cautions. Measure engagement at team level with a minimum cell size and publish that minimum before you field the survey, or the honest answers will not arrive. And do not run the engagement instrument inside a performance review cycle, for the same reason.
Culture problem or management-system problem?
This is the question worth asking before committing a budget, because the two have different fixes and the wrong diagnosis is expensive. A management-system problem responds to a process change in months. A genuine culture problem does not, and treating one as the other wastes a year.
| What you observe | More likely a management-system problem | More likely a culture problem |
|---|---|---|
| Where the weakness sits | Concentrated in specific teams, functions or sites | Consistent across every part of the organisation |
| What people say when asked privately | They can describe the specific blocker and who owns it | They describe the behaviour as normal and do not see it as a problem |
| What happens when a rule changes | Behaviour follows within a quarter | Behaviour reverts once attention moves elsewhere |
| New joiners | Arrive and are frustrated by the constraint | Arrive, comment on it for a month, then adopt it |
| The gap between stated and actual | People acknowledge it openly | People do not perceive a gap at all |
Most enterprise situations turn out to be predominantly management-system problems wearing culture language, which is good news, because those are the ones that move. The tell is the fourth row. If new joiners stop noticing within a month, the environment is teaching them something, and the something is worth finding.
Where MASSIVUE fits
MASSIVUE is an enterprise transformation and AI consulting firm headquartered in Singapore, working with organisations across financial services, insurance, retail, maritime and technology. We combine consulting with certified training delivered through MASSIVUE Academy, on the view that capability which leaves when the consultants leave was never capability.
On this problem specifically, three parts of our work are directly relevant:
- Enterprise Transformation covers stages 1, 3 and 5: baselining where performance actually varies, confronting the reward system, and codifying what the organisation has proved it can do.
- AI Workforce Transformation covers stages 2 and 4: building the manager and workforce capability, and redesigning the work rather than layering tools onto it.
- MASSIVUE Academy supplies the training component, so that stage 2 is a standing capability with named ownership rather than a one-off vendor course.
What is MASSIVUE-owned in this article and what is not: the five-stage sequence, the owner-and-observable structure, the measurement set, and the culture-versus-management-system decision table are ours, developed through client transformation work. Every statistic is externally sourced and attributed below, and where a claim is our judgement rather than a research finding, it is labelled at the point of use.
If you would rather start with the diagnosis, book a 30-minute chat. We will look at where your team-level variance actually sits and tell you plainly which of the five stages you are missing.
Frequently asked questions
What is a high-performance culture?
A high-performance culture is a pattern of behaviour, repeated across teams, in which people consistently do the work that produces results without needing to be told. It is observable in what happens under pressure and when nobody is supervising. It is distinct from climate, which is how the environment feels at a given moment, and from engagement, which measures discretionary effort and intent to stay.
How long does it take to build a high-performance culture?
Manager engagement typically moves first, within one to two quarters of real investment in the manager layer. Team engagement follows over two to four quarters. Business outcomes are a year-plus horizon. Anything promising a cultural change within a quarter is describing a climate change, which is real but reverts when attention moves elsewhere.
Do values statements matter at all?
They matter as a record of behaviour you have already produced, not as an instruction for behaviour you have not. Across 562 companies that published values statements, researchers found no correlation between the values emphasised and how employees rated the company on those same values, with four of nine values correlating negatively. Publish them last, and only where you can point at a specific instance from the last quarter.
Is employee engagement the same as a high-performance culture?
No, but it is the best-evidenced proxy available. Engagement measures discretionary effort and intent to stay; culture is the underlying pattern of behaviour that produces it. Engagement is useful because it is measured at team level and has a documented link to eleven business outcomes across 183,806 business units, which makes it the practical instrument even though it is not the thing itself.
Why is manager engagement falling?
Gallup points to a combination of expanding spans of control, restructuring, and the removal of layers, some of it driven by AI adoption. Manager engagement declines as span of control grows, though Gallup finds manager talent and training can offset that effect. The practical implication is that if you are flattening the organisation, the manager development budget needs to rise rather than fall.
Where does AI fit into a culture roadmap?
At stage 4, and not as a separate programme. AI has reached individuals without reaching the operating model: 65% of workers in AI-adopting organisations report a personal productivity gain, only 12% strongly agree it has transformed how work gets done, and 89% of executives report no measurable firm-level productivity effect over three years. That pattern is a management-system problem, so it belongs inside the same roadmap rather than alongside it.
What is the single cheapest thing to do first?
Find out how many of your people managers have ever received role training, and how many have a named person responsible for their development. Globally, 44% of managers have had any management training. Trained managers are half as likely to be actively disengaged, and manager thriving rises from 28% to 34% with training and to 50% when someone actively encourages their development.
Related reading
- How Do You Build an AI-Ready Culture? Start With What Your People Are Hiding. This article covers how to build performance culture generally. That one goes deep on a single measurable culture signal, the gap between the AI use that happens and the AI use people will admit to, and the five conditions that close it.
- End-to-End Change Management in Singapore: A 2026 Playbook. If your immediate problem is landing a specific change rather than raising the baseline, start there.
- The Silent Productivity Killer: Understanding and Measuring the Impact of Toxic Workplace Culture. The same measurement logic applied to the downside case.
Sources
- James K. Harter, Corey E. Tatel, Sangeeta Agrawal, Anthony Blue, Stephanie K. Plowman, Jim Asplund, Sabrina Yu and Andy Kemp, The Relationship Between Engagement at Work and Organizational Outcomes: Q12 Meta-Analysis, 11th Edition, Gallup, May 2024, revised July 2024 (736 studies, 347 organisations, 53 industries, 90 countries, 183,806 business units, 3,354,784 employees). gallup.com
- Gallup, State of the Global Workplace: 2026 Report (Gallup World Poll; 2025 fieldwork January to December 2025, 263,810 respondents of whom 141,444 were employed, 140+ countries and territories). gallup.com
- Gallup, State of the Global Workplace: 2025 Report (manager training prevalence and effects). gallup.com
- Randall J. Beck and Jim Harter, Why Great Managers Are So Rare, Gallup Business Journal, 25 March 2014, updated 16 February 2026 (source of the 70% variance estimate and the 82% selection figure). gallup.com
- Ivan Yotzov, Jose Maria Barrero, Nicholas Bloom, Philip Bunn, Steven J. Davis, Kevin M. Foster, Aaron Jalca, Brent H. Meyer, Paul Mizen, Michael A. Navarrete, Pawel Smietanka, Gregory Thwaites and Ben Zhe Wang, Firm Data on AI, NBER Working Paper 34836, February 2026, revised March 2026 (nearly 6,000 senior executives in the US, UK, Germany and Australia, fielded November 2025 to January 2026 with the Federal Reserve Bank of Atlanta, Bank of England, Deutsche Bundesbank and Macquarie University). nber.org
- Donald Sull, Stefano Turconi and Charles Sull, When It Comes to Culture, Does Your Company Walk the Talk?, MIT Sloan Management Review, 21 July 2020 (689 large companies analysed, 562 with published values statements, 1.2 million Glassdoor reviews). sloanreview.mit.edu