Building performance management from scratch: the Revolut Playbook

Three dimensions that cannot be reduced to one KPI

The Revolut Playbook separates performance into deliverables, skills and culture. Deliverables are the ability to take a task through to a result. Speed is assessed against the agreed deadline, quality by the amount of rework needed, and complexity by whether the task had a precedent. In this model, the components are standardised and the outcome is calculated as (speed + quality) × complexity. A new task with no ready-made solution cannot be assessed in the same way as a routine operation.

Skills are specific professional capabilities required for a role. For a product manager, these include problem definition, project management and commercial thinking; an engineer has different criteria. The head of the relevant function defines them.

Culture is not a test of personality similarity. A value is translated into observable behaviour: does the person take responsibility, flag risks early and help an adjacent team? Words such as “initiative” and “teamwork”, without examples, return assessment to a manager’s impressions.

Team KPIs and an individual’s assessment are therefore not the same thing. A department may have exceeded its plan thanks to a favourable market; a manager may have run a difficult experiment well without achieving the expected effect. The CEO needs to see both: the business result and the professional work behind it.

How a broad expectation becomes a verifiable criterion

For each dimension, the Revolut Playbook proposes a scorecard: a description of what someone should be able to do and demonstrate at five levels — Poor, Basic, Intermediate, Advanced and Exceptional. The manager answers “yes” or “no” to statements about specific actions. Moving up requires confirmation of every requirement at the previous level; the first unconfirmed statement marks the limit of demonstrated proficiency.

For a data skill, Basic might mean reading an existing report; Intermediate, choosing a metric and checking its source; Advanced, designing an experiment and explaining the limits of its conclusions. The manager adds an example of work to the answer. This is how a scorecard becomes a development tool.

Talent bar: why the same rating means different things for different roles

The talent bar is the minimum acceptable level for each dimension, taking the role and seniority into account. Skills vary by role; requirements for deliverables and culture may be shared across the company. A Junior carries out defined work with support; a Senior independently solves more complex problems and influences other people’s work.

Now return to Alice, Chris and Mary. Alice exceeds the bar almost everywhere — she is a candidate for greater responsibility. Chris meets the bar: this is a good outcome for his role, not “mediocrity”. Mary is below the bar in at least one dimension; in this model, strong deliverables do not compensate for a failure in culture. This illustrates a principle the CEO must define before launching assessment: which requirements are non-negotiable for the company?

The system needs an owner, not just a form

The Revolut Playbook calls performance a direct mandate of the CEO. A dedicated performance team owns the rules, data and cycle calendar: it maintains role matrices, checks KPI quality, collects reviews, calculates and calibrates results, and prepares reward decisions. It does not replace the functional leader, who sets the team’s goals and assesses people’s work. HR ensures a sound process and supports employees.

For a company of up to 100 people, one strong operations manager alongside the CEO is enough. They can check where a metric came from, who influences it and whether managers share the same understanding of Senior level. As the company grows, this role can develop into a separate function.

What actually happens during a quarterly review

The Revolut Playbook cycle has three stages. For 1–2 weeks, direct managers complete scorecards based on observed work; other managers can provide context but do not replace the person responsible for the assessment. Over the following week, the performance team checks calculations, omissions and anomalies. Why has half the department suddenly become A-players? What supports a Junior’s jump to Exceptional? Have different people received different ratings for the same actions? After calibration, the CEO and senior management approve the outcome. A further week is allocated to the manager’s one-to-one conversation with each person.

The Revolut Playbook model limits A-players to 15–25% of the team — a deliberate choice to set a very high bar. Calibration, however, should not become a negotiation over quotas: if half a function has exceeded expectations, the CEO checks the evidence, the height of the bar and the context of the work.

Assessment must change decisions

In this system, strong performance has consequences for careers and rewards. Promotions are considered every six months: they require at least a year at the current level, no below-bar ratings and a case for entrusting the person with more complex work. For a move from Junior to Mid, the Revolut Playbook specifies at least two A-player ratings in the last four reviews; for moves above Mid, four out of four. A new job title alone is not development.

Rewards are tied to the team’s or company’s results as well as the individual’s rating. Annual bonus = potential bonus × weighted multiplier for individual and team performance × proportion of the period worked. The more senior the leadership role, the greater the weight of collective performance: in the Revolut Playbook example, it is 20% for a Junior and 55% for a Director. This makes sense: a leader should also be assessed on the system they have built for others. If someone is below the bar, the model proposes a six-week improvement plan with written goals, an interim check and a final decision.

How to launch the system in 90 days

I would start with one function rather than the whole company. The first two weeks: the CEO and functional leader define the business outcome, the team’s area of accountability and the decisions it genuinely influences. The next two: they describe two or three key roles, the talent bar and short scorecards with examples of behaviour. The next six: they hold weekly one-to-ones, record agreements and gather evidence of work. The final three: they conduct a trial review, calibrate disputed cases and check whether people understand the outcome.

Before linking assessment to money, ask two managers to rate the same cases independently. A discrepancy will reveal an unclear bar, missing evidence or differing standards. Resolve this before making career and financial decisions.

The CEO’s central question at the end of the pilot is whether each person can explain what result is expected of them, how the complexity of their work will be assessed and what will unlock their next level of responsibility. If the answer is clear, the system has already started working.