What you measure is what you become
How metric trees shape company culture
Metrics are not neutral. The structure of your measurement system sends constant signals about what your organisation values, how teams should relate to one another, and whether people should hide problems or surface them. A metric tree, designed well, becomes a cultural artefact that reinforces collaboration, transparency, and learning. Designed poorly, it breeds fear, gaming, and silos.
9 min read
Metrics as cultural artefacts
Every organisation has a set of official values printed on the wall, and a set of real values revealed by how people behave under pressure. The real values are almost always shaped by what gets measured. When a company tracks individual sales targets but not customer satisfaction, it is telling salespeople that closing deals matters more than keeping customers happy, regardless of what the values poster says about being "customer-first." When an engineering team is measured on velocity but not reliability, shipping speed will win every trade-off against code quality, no matter how many times the CTO talks about craftsmanship at all-hands meetings.
This is not a failure of character. It is basic behavioural economics. People respond to incentives, and metrics create incentives whether you intend them to or not. The metrics you choose to track, the ones you review in meetings, the ones you tie to performance evaluations, these are the most honest expression of your organisational priorities. They function as cultural artefacts: tangible, visible objects that encode and transmit the beliefs, values, and assumptions of the group. Anthropologists study pottery and tools to understand ancient civilisations. If you want to understand a modern company, look at its metrics.
The cultural mirror
If you want to know what your organisation truly values, ignore the mission statement and look at the metrics. What gets measured gets managed, but more importantly, what gets measured gets signalled. Every metric you track tells your people: "This is what matters here." Choose carefully, because your team is listening.
A metric tree makes this dynamic explicit rather than accidental. By organising metrics into a visible hierarchy, it reveals the full picture of what the organisation prioritises and, equally importantly, what it does not. When a team can see that their metric connects upward to the company North Star, they understand that their work matters. When they can see that their metric sits alongside complementary metrics from other teams, they understand that success is shared. The tree does not just track performance. It communicates purpose, context, and interdependence. It is the organisational chart of what matters.
How tree structure promotes collaboration over competition
One of the most consequential design choices in any measurement system is whether it encourages collaboration or competition between teams. Many organisations default to competition without realising it. When every team has isolated KPIs with no visible connection to each other or to a shared outcome, each team optimises locally. Marketing optimises for leads. Sales optimises for closed deals. Product optimises for feature releases. Each team hits its numbers, but the overall business outcome stagnates or declines because nobody owns the connections between the metrics.
A metric tree solves this by making the connections visible and structural. When marketing can see that their leads metric feeds into a qualified pipeline metric that feeds into revenue, they stop optimising for volume and start optimising for quality. When product can see that their feature adoption metric sits alongside customer support ticket volume under a shared retention node, they think twice before shipping half-finished features that create support burden. The tree reveals that optimising one metric in isolation can damage a sibling metric, and that the real goal is improving the parent. This shared visibility transforms the incentive structure from competitive to collaborative.
Notice how the tree above reveals shared ownership. Trial-to-paid conversion is not just a sales metric or a product metric. It depends on both teams working together. Expansion revenue requires collaboration between sales and customer success. When the tree makes these dependencies visible, cross-functional collaboration becomes the obvious strategy rather than something leadership has to mandate. Teams collaborate not because they are told to, but because the structure shows them that their success depends on it. The best collaboration does not come from team-building exercises. It comes from shared metrics with visible interdependencies.
Transparency and psychological safety
Metric transparency, making performance data visible across the organisation, is a necessary condition for a healthy measurement culture. But transparency alone is not sufficient. Without psychological safety, transparency becomes surveillance. People who fear punishment for bad numbers will hide problems, massage data, delay reporting, or simply stop looking at the metrics altogether. The combination of high visibility and low safety is one of the most toxic cultural patterns an organisation can create.
| Dimension | High safety + transparency | Low safety + transparency |
|---|---|---|
| When a metric drops | Team surfaces the issue early, investigates root causes openly, and shares findings across the organisation | Team delays reporting, looks for external explanations, or adjusts the metric definition to make the number look better |
| When a target is missed | Honest post-mortem focused on what was learned and what to change next time | Blame assignment, defensive presentations, and quiet removal of the target from future reporting |
| When teams disagree on data | Productive debate about definitions, methodology, and interpretation that strengthens shared understanding | Political manoeuvring where each team promotes the numbers that make them look best |
| When an experiment fails | Result is documented and shared as institutional knowledge that prevents repeating the same mistake | Result is buried, and the experiment is reframed as a success using cherry-picked secondary metrics |
A well-designed metric tree supports psychological safety in two important ways. First, it distributes accountability across the hierarchy. When a high-level metric drops, the tree structure shows that the cause could be anywhere in the branches below it. This shifts the conversation from "who is to blame?" to "where in the tree did the breakdown occur?" The investigation becomes structural rather than personal. Second, the tree normalises metric movement. When everyone can see that metrics naturally fluctuate, that some branches improve while others decline, and that this is how complex systems behave, people stop treating every dip as a crisis and every spike as a triumph. The tree teaches the organisation to think in systems rather than in snapshots.
“The purpose of making metrics visible is not to create pressure. It is to create a shared understanding of reality. When people trust that honest numbers will be met with curiosity rather than punishment, they stop hiding problems and start solving them.”
Ownership without blame
Metric ownership is essential. Without a named person responsible for understanding why a metric moves and what to do about it, metrics become decorative. But there is a critical distinction between ownership and blame that many organisations fail to maintain. Ownership means "I am the person who will investigate this metric, understand its behaviour, and propose actions." Blame means "I am the person who will be punished when this metric goes wrong." The first drives engagement. The second drives fear.
The structure of a metric tree helps maintain this distinction because it makes the systemic nature of performance visible. When a metric owner can point to the tree and show that their metric declined because an upstream input changed, or because a sibling metric was prioritised at the expense of theirs, the conversation stays grounded in the system rather than collapsing into personal accountability for outcomes that no single person controls. The tree provides the context that prevents ownership from degenerating into blame.
- 1
Separate understanding from outcomes
Hold metric owners accountable for understanding their metric deeply and acting on what they learn, not for hitting a specific number. Many factors that influence a metric are outside any individual's control. What is within their control is whether they are paying attention, investigating, and responding intelligently.
- 2
Celebrate diagnostic skill
When a metric owner identifies the root cause of a movement quickly and accurately, recognise that as a win, even if the metric itself moved in the wrong direction. The ability to diagnose is more valuable in the long run than the ability to hit a target through luck or favourable conditions.
- 3
Use the tree to contextualise movement
When reviewing metrics, always look at the surrounding branches. A decline in one metric that coincides with growth in a related metric might represent a healthy trade-off, not a failure. The tree gives you the vocabulary to discuss these trade-offs without defaulting to blame.
- 4
Make ownership rotational where appropriate
For some metrics, rotating ownership periodically prevents it from becoming a permanent burden that people associate with punishment. Rotation also builds organisational resilience, as multiple people develop deep familiarity with different parts of the tree.
- 5
Distinguish between owner and fixer
The metric owner is the person who understands and reports on the metric. They are not necessarily the person who fixes it when something goes wrong. The tree often reveals that the fix lives in a different branch from the symptom. Clarifying this distinction protects owners from being held responsible for problems they can diagnose but not directly solve.
When the wrong metrics create toxic culture
Goodhart's Law states that when a measure becomes a target, it ceases to be a good measure. Campbell's Law extends this further: the more a quantitative indicator is used for decision-making, the more it will be subject to corruption pressures and the more apt it will be to distort and corrupt the processes it is intended to monitor. These are not abstract academic principles. They describe the lived reality of organisations where poorly chosen metrics have warped behaviour, eroded trust, and created cultures that nobody wanted but everybody contributed to.
The speed trap
Measuring engineering teams purely on velocity or story points delivered incentivises teams to inflate estimates, split tickets artificially, and ship features without adequate testing. The metric goes up while the product gets worse. Teams learn that appearing productive is rewarded more than being productive.
The volume trap
Measuring sales on activity volume, calls made, emails sent, meetings booked, creates a culture of busywork where the appearance of effort replaces the pursuit of outcomes. Salespeople optimise for the metric rather than the customer, and pipeline quality deteriorates while activity dashboards glow green.
The utilisation trap
Measuring team utilisation rates above 90% signals that busyness is valued over impact. Teams lose time for learning, reflection, and strategic thinking. People pad their timesheets and avoid helping colleagues because time spent on someone else's project does not count toward their utilisation target.
The satisfaction trap
Measuring customer satisfaction through post-interaction surveys incentivises staff to ask for high ratings, cherry-pick which customers receive surveys, or resolve easy tickets first and escalate difficult ones. The satisfaction score rises while actual service quality stagnates or declines.
The common thread in all of these traps is that the metric was chosen without considering the behaviours it would incentivise. A metric tree helps prevent these failures in three ways. First, balance: because the tree includes metrics across multiple dimensions, it is much harder to game one metric without visibly damaging another. If an engineering team inflates velocity, the tree will show corresponding declines in reliability or customer satisfaction in sibling branches. Second, context: the tree shows what a metric is supposed to drive, making it easier to spot when the metric is improving but the outcome it serves is not. Third, conversation: the tree creates a natural forum for discussing whether metrics are producing the right behaviours, because the relationships between metrics make distortions visible rather than hidden.
The antidote to gaming
You cannot prevent gaming by adding more metrics. You prevent it by making the system of metrics visible enough that gaming one number at the expense of others becomes obvious. A metric tree does this structurally. When every metric has siblings, parents, and children, optimising one at the expense of the system is no longer invisible. It shows up in the tree.
Using the tree to reinforce values
If metrics are cultural artefacts, then designing your metric tree is an act of cultural design. Every choice you make about what to include, what to exclude, how to structure the hierarchy, and where to place ownership is a statement about what your organisation values. This means that building a metric tree is not purely a data exercise. It is a leadership exercise that deserves the same thoughtfulness you would apply to defining your company values, because the tree will have a far greater impact on daily behaviour than any values statement ever will.
- 1
If you value collaboration, measure shared outcomes
Place metrics at the intersection of teams rather than within silos. Revenue is not a sales metric. It is a company metric that depends on marketing, product, sales, and customer success working together. Position it in the tree so that the contributing branches are visible, and teams will naturally coordinate.
- 2
If you value quality, measure it alongside speed
Never measure throughput without a corresponding quality metric in the same branch of the tree. If you measure features shipped, also measure defect rates. If you measure tickets closed, also measure reopen rates. The tree should make it structurally impossible to celebrate speed without accounting for the quality of what was delivered.
- 3
If you value learning, measure experiments and hypotheses
Include metrics that track how many hypotheses the organisation is testing, not just how many succeed. A tree that only tracks outcomes tells people that results are all that matter. A tree that also tracks experimentation velocity tells people that the process of learning is valued, even when individual experiments fail.
- 4
If you value customer focus, let the customer appear in every branch
Ensure that customer-facing metrics appear at every level of the tree, not just in a single "customer" branch. When product teams, engineering teams, and operations teams can all see how their work connects to customer outcomes, customer focus becomes systemic rather than the responsibility of a single department.
- 5
If you value sustainability, measure leading and lagging together
Pair short-term output metrics with long-term health indicators in the same tree. Revenue growth alongside employee engagement. Feature velocity alongside technical debt. Customer acquisition alongside retention. The tree should make it impossible to optimise for the present at the expense of the future without the trade-off being visible.
The most powerful cultural interventions are often the simplest structural changes to the tree. Adding a quality metric as a sibling to a speed metric changes behaviour overnight. Moving a metric from one branch to another redefines who cares about it and who collaborates on it. Removing a metric that has been driving perverse incentives sends a clear signal that the organisation has learned and adapted. The tree is a living document of cultural intent, and every edit to it is a cultural decision.
“Your metric tree is the most honest expression of your company culture. Not what you say you value, but what you actually track, review, and reward. If you do not like the culture you have, look at the tree you have built. The culture will not change until the tree does.”
Designing a metric tree for the culture you want
Building a metric tree that reinforces healthy culture is not something you do once and forget. It is an ongoing practice of observation, reflection, and adjustment. The tree will inevitably produce unintended consequences, because all measurement systems do. The difference between organisations that build great cultures and those that accidentally destroy them is whether they notice those consequences and respond. Here is a practical framework for keeping your tree culturally aligned.
Audit for perverse incentives
Quarterly, review each metric in the tree and ask: "If a team optimised exclusively for this number, what behaviour would that produce?" If the answer describes behaviour that conflicts with your values, the metric needs a counterbalancing sibling or a redesign.
Involve the people being measured
The teams whose behaviour a metric is meant to influence should have a voice in choosing and refining that metric. Imposed metrics breed resentment. Co-created metrics build buy-in. When people help design their own measurement, they are more likely to engage with it honestly.
Review the shape of the tree
A tree that is deep and narrow suggests an organisation focused on a single dimension of performance. A tree that is broad and shallow suggests one that is tracking everything but connecting nothing. The shape of your tree should reflect the balanced, interconnected nature of your business.
Protect the guardrail metrics
Some metrics exist not to be optimised but to ensure that optimising other metrics does not cause harm. Employee wellbeing, customer satisfaction floors, and quality thresholds are guardrails. Make sure they are visible in the tree and that they cannot be quietly deprioritised.
Evolve the tree as the culture matures
A young organisation might need more output metrics to build momentum. A mature organisation might need more learning metrics to prevent stagnation. The tree should evolve with the organisation, reflecting not just what you need to achieve today but the kind of company you are becoming.
Culture is not what you declare. It is what you do repeatedly. And what you do repeatedly is shaped, more than most leaders realise, by the metrics you choose to make visible, the structure you give them, and the conversations those metrics enable. A metric tree is not just a performance management tool. It is a cultural architecture. Build it with the same care and intentionality you would bring to any other foundational decision about the kind of organisation you want to be.
Build a metric tree that reinforces the culture you want
Your metrics shape behaviour whether you design them to or not. Map your metric tree with KPI Tree to make the connections visible, distribute ownership without blame, and create a measurement system that reinforces collaboration, transparency, and learning.
