I was recently asked to be part of a focus group for a new performance management framework. One of the things I noticed is that the framework fell into the same trap as what I’ve seen at other companies I’ve worked that had the same framework. I’m not sure where these HR teams get the ideas for these frameworks, but they all suffer from the same flaws. Regardless, I wanted to highlight an issue I saw with their 9-Box model.

The 9-Box was being measured on two axes: impact and growth velocity. The impact axis I had no issue with. The growth velocity, however, has the innate problem in that it is not measurable as defined. At a previous company, they called it growth potential. I raised the same objections then but was dismissed. Now, I’m at a more open-minded company and I was asked for what would “good” look like. It gave me some pause and then I had to admit that I needed more time, that this topic would be good to discuss off-line. The reason is that fixing the measurability of a person’s growth is backed by complexity, so first understanding that complexity makes it then simple to apply.

This page is my attempt to bring measurement to the growth debate. It will never be 100% accurate but it will be evidence-based, which helps to mitigate biases and subjectivity. The end goal is to move the measurement of someone’s growth out of the realm of vibes, and into the realm of science.

It is borne out of how I’ve always wanted my growth to be measured.

This is now an interactive tool: Growth Velocity Rubric or check the Tools link at the top of this page.

Growth Velocity Rubric Specification

Purpose

The Growth Velocity Rubric is a structured assessment tool for evaluating how quickly and effectively an employee is becoming more capable over time.

It is designed to support performance management, talent reviews, calibration discussions, and 9-box-style assessments where “growth,” “potential,” or “trajectory” may otherwise be evaluated subjectively.

The core problem this rubric solves is that growth velocity is often discussed as if it is obvious, but without a rubric it can easily become based on vague impressions, charisma, self-promotion, manager preference, or recency bias.

This rubric provides:

• A consistent scoring model.
• Role-normalized expectations.
• Observable behavioural criteria.
• Evidence capture.
• A low / medium / high growth velocity rating.
• Confidence scoring.
• A generated summary for calibration or documentation.

The rubric is not intended to replace manager judgment. It is intended to discipline manager judgment so that ratings are based on observable evidence rather than unsupported impressions.

Definition of Growth Velocity

We need a definition to intentionally focus on visible change over time. Growth velocity is defined as:

The rate at which someone converts feedback, new challenges, and increasing ambiguity into demonstrably better performance.

Growth velocity is not the same as:

• Effort.
• Ambition.
• Confidence.
• Busyness.
• Working long hours.
• Taking training courses.
• Being well-liked.
• Having high impact today.
• Being early-career and therefore learning many visible basics.

Growth velocity asks:

Is this person becoming more capable, and is that increased capability showing up in their work?

Relationship Between Impact and Growth Velocity

Impact and growth velocity are related but distinct. The first is backwards-looking. The second is forward-looking. Anything based in the future is naturally prone to a lack of measurement.

Impact asks:

What did the person deliver?

Growth velocity asks:

How much more capable is the person becoming, and how quickly is that change appearing in their work?

For example, a person may have high impact but medium growth velocity if they continue delivering well but are not materially expanding capability, judgment, independence, or scope.

Another person may have medium current impact but high growth velocity if they are rapidly improving, taking feedback well, expanding scope, and becoming effective in more ambiguous work.

The rubric is meant to make this distinction clearer during performance calibration.

Assessment Dimensions

The rubric evaluates seven dimensions:

1. Feedback uptake
2. Scope expansion
3. Skill acquisition
4. Judgment improvement
5. Independence
6. Learning transfer
7. Time compression

Each dimension is scored from 0 to 3.

The maximum total score is 21.

Limitations

The rubric has limitations.

• It does not fully remove subjectivity. It structures subjectivity.
• It does not prove growth velocity completely scientifically. It creates a more consistent evidence-based assessment.
• It may still reflect bias if managers provide biased evidence.
• It may be harder to use when managers have weak observation quality.
• It may require calibration to avoid score inflation.
• It may need job-family examples (see Appendix) to become operationally useful across a whole company.
• It should not be used as the only input into compensation, promotion, or succession decisions.

Scoring Scale

Each dimension uses the following general scoring pattern:

A score of 2 should represent good, expected growth for the selected role level.

A score of 3 should be reserved for growth that is faster, broader, more repeatable, or more leveraged than expected.

The scoring model deliberately uses a small 0–3 scale because growth velocity is not precise enough to justify fine-grained scoring. A larger scale would create false precision. But scoring is just one facet and it can cause more experienced people to look like they have less velocity than less experienced people, simply because less experienced people have more headroom.

Role Normalization

Role normalization is central to the rubric.

The same growth dimension means different things at different levels. For example, “independence” for an early-career employee may mean completing familiar tasks with less guidance. For a senior/staff employee, independence may mean turning ambiguous organizational problems into clear direction for others.

Role normalization is applied inside each dimension score.

The tool does not calculate a raw score and then adjust it after the fact. Instead, the selected role level determines which behavioural expectations should be used when assigning each score.

The supported role levels are:

1. Junior / Early Career
2. Intermediate
3. Senior / Staff

The selected role level is highlighted in the rubric table so the evaluator can compare the employee against the appropriate expectation set.

The final score therefore means:

Total growth velocity score against the expected growth path for the selected role level.

Role-Level Philosophy

A 3-level philosophy is used. It is possible to add more, but realize that more stratification doesn’t always lend itself to more accuracy.

Junior / Early Career

At this level, growth is often visible through:

• Learning foundational skills.
• Becoming reliable on familiar tasks.
• Needing less direct guidance.
• Applying feedback to day-to-day work.
• Avoiding repeat mistakes.
• Moving from task execution toward small workstream ownership.

High growth velocity at this level usually means the person is becoming productive and reliable faster than expected.

Intermediate

At this level, growth is often visible through:

• Owning normal role-scope work independently.
• Expanding from tasks to projects, workflows, or features.
• Applying judgment to tradeoffs.
• Transferring learning across projects.
• Becoming effective in new responsibilities within expected time.
• Reducing the need for manager intervention.

High growth velocity at this level usually means the person is expanding ownership and decision quality faster than expected.

Senior / Staff

At this level, growth is often visible through:

• Improving judgment in ambiguous, high-impact, or cross-functional situations.
• Creating clarity where there is ambiguity.
• Expanding influence across teams or systems.
• Translating learning into standards, patterns, practices, coaching, or reusable frameworks.
• Improving the work of others, not only their own output.
• Building organizational leverage.

High growth velocity at this level is less about learning more visible basics and more about increasing leverage, influence, judgment, and system-level contribution.

Dimension Criteria

These are a set of attributes to gauge someone’s growth. Because growth is not just a simple metric, and we want to disincentivize burnout, we focus these criteria across different dimensions.

Feedback Uptake

Feedback uptake measures whether the employee receives feedback and converts it into changed behaviour or improved work.

Guidance

Feedback uptake should be based on observable change, not whether the person seemed receptive in the conversation. Someone can appear gracious when receiving feedback but not change behaviour. Conversely, someone may need time to process feedback but later apply it well.

Useful evidence includes:

• Reduced repeat issues.
• Better stakeholder communication.
• Improved work quality after feedback.
• Peer or stakeholder confirmation.
• Clear before/after examples.

Scope Expansion

Scope expansion measures whether the employee is growing into broader, more complex, or more ambiguous ownership.

Guidance

Scope expansion should not mean simply giving someone more work. It should mean the person is successfully operating at a broader level of ownership or complexity.

For senior/staff employees, scope expansion should often involve influence beyond a single task or project. This might include cross-team standards, reusable practices, architectural direction, operational improvements, or system-level change.

Skill Acquisition

Skill acquisition measures whether the employee is developing new capabilities and applying them in meaningful work.

Guidance

Skill acquisition should not be scored based only on completed training, certifications, or courses. The important question is whether the new skill appears in delivered work, better decisions, better execution, or increased team capability.

For senior/staff employees, learning a tool may not be sufficient evidence of high growth velocity. Stronger evidence would be using an adjacent skill to improve team practices, architecture, security, delivery quality, customer outcomes, or decision-making.

Judgment Improvement

Judgment improvement measures whether the employee is making better decisions, identifying risks earlier, and improving tradeoff quality.

Guidance

Judgment improvement is one of the most important dimensions for senior employees. At higher levels, growth may be less visible as task execution and more visible as improved prioritization, risk detection, tradeoff quality, and decision framing.

Useful evidence includes:

• Fewer preventable surprises.
• Earlier escalation of meaningful risks.
• Better prioritization.
• Stronger recommendations.
• Improved architecture, design, operational, or stakeholder decisions.
• Helping others make better decisions.

Independence

Independence measures whether the employee needs less direction over time and can operate effectively with appropriate autonomy for their level.

Guidance

Independence does not mean never asking questions or never escalating issues. Healthy independence includes knowing when to ask, when to decide, when to escalate, and when to create clarity for others.

For senior/staff employees, the strongest evidence is not merely that they can work alone. It is that they can take ambiguous situations and turn them into direction, structure, decisions, or reusable approaches.

Learning Transfer

Learning transfer measures whether the employee applies lessons from one situation to other contexts.

Guidance

Learning transfer is especially important because growth velocity should not only measure isolated improvement. It should measure whether the person can generalize what they learn.

For senior/staff employees, learning transfer should often produce leverage. Examples include:

• Standards.
• Playbooks.
• Design patterns.
• Decision records.
• Coaching.
• Improved defaults.
• Reusable frameworks.
• Process improvements.
• Reduced future risk.

Time Compression

Time compression measures how quickly the employee reaches effective capability after exposure to new work, new context, or new ambiguity.

Guidance

Time compression should be used carefully. It should not reward overwork, unsustainable intensity, or constant urgency.

The dimension measures speed to effective capability, not speed to exhaustion.

Evidence should show that the person became genuinely effective faster than expected, not merely that they worked longer hours or appeared busy.

Evidence Model

Each dimension includes three evidence fields:

1. Before
2. After
3. Evidence

Before

This captures the prior state.

Examples:

• Needed repeated reminders to update stakeholders.
• Required detailed task breakdowns.
• Missed dependencies until late in delivery.
• Had limited understanding of a domain.
• Struggled to apply feedback across contexts.

After

This captures the observable change.

Examples:

• Now sends concise weekly stakeholder updates without prompting.
• Breaks down ambiguous work independently.
• Identifies cross-team dependencies earlier.
• Applies new domain knowledge to improve design choices.
• Uses feedback from one situation in another project.

Evidence

This captures the artifact, observation, or validation that supports the score.

Examples:

• Delivered project.
• Design document.
• Decision record.
• Peer feedback.
• Stakeholder feedback.
• Manager observation.
• Incident review.
• Sprint output.
• Architecture review.
• Customer outcome.
• Reduced repeat issues.
• New standard, playbook, or reusable pattern.

The evidence model exists because a numeric score without evidence becomes subjective. The rubric should require evidence for any non-zero score, especially scores of 2 or 3.

Total Score and Rating Bands

Each of the seven dimensions is scored from 0 to 3.

Maximum score:

7 dimensions × 3 points = 21 points

The rating bands are:

Low Growth Velocity

A low rating means evidence is mostly absent, inconsistent, or below current role-level expectations.

This does not necessarily mean the person is low-performing overall. It means the evidence does not show meaningful growth velocity during the review period.

Medium Growth Velocity

A medium rating means evidence shows normal, healthy growth for the selected role level.

This should be considered a positive and acceptable rating. Most people should likely be in this band most of the time if they are growing appropriately.

High Growth Velocity

A high rating means evidence shows growth beyond current role-level expectations across multiple dimensions.

A high rating should be uncommon and evidence-heavy.

High-Rating Gate

To avoid inflated high ratings, the rubric includes a high-rating gate.

For a defensible High rating, the assessment should meet both of the following conditions:

1. At least 4 of the 7 dimensions are scored 2 or higher.
2. At least 2 dimensions are scored 3.

This prevents a person from being rated High based on one exceptional area or a single recent success. If the total score reaches the High band but the gate is not satisfied, a warning should be displayed so that there is a chance to understand the improvement path. This does not prevent the evaluator from seeing the numeric score, but it signals that the High rating may not be well-supported.

Confidence Rating

The rubric also needs a separate confidence rating:

• Low
• Medium
• High

Confidence is not part of the numeric score. It describes how strong the evidence base is.

Low Confidence

Use Low confidence when:

• Evidence is mostly anecdotal.
• There are too few examples.
• The before/after state is unclear.
• The rating is based on a narrow interaction.
• There is little stakeholder or artifact support.

Medium Confidence

Use Medium confidence when:

• There are some concrete examples.
• Evidence is real but limited.
• Evidence may be concentrated in one project.
• More time or broader validation would improve confidence.

High Confidence

Use High confidence when:

• There are multiple examples across time.
• Evidence comes from more than one project, stakeholder, or artifact.
• The before/after pattern is clear.
• The growth appears repeatable.
• The rating would likely survive calibration.

The confidence rating is important because someone may appear to have high growth velocity based on a single project, but the evidence may not yet be broad enough to support high confidence.

Intended Usage

The rubric can be used by:

• Managers preparing performance reviews.
• P&C teams designing performance frameworks.
• Calibration groups.
• Talent review committees.
• Employees preparing self-assessments.
• Mentors or career coaches helping someone understand growth areas.

Recommended usage:

1. Select the employee’s current role level.
2. Review each dimension.
3. Score against the selected role-level criteria.
4. Capture before/after/evidence notes.
5. Review the total score.
6. Check whether the high-rating gate passed, if applicable.
7. Set evidence confidence.
8. Use the generated summary for calibration or documentation.

Calibration Guidance

During calibration, reviewers should ask:

• What changed over the review period?
• Is the change observable?
• Is the evidence tied to work output, behaviour, or stakeholder impact?
• Is the person being scored against the correct role level?
• Are we rewarding effort, or actual capability growth?
• Are we rewarding confidence or communication polish instead of evidence?
• Is the rating based on one recent event?
• Would another manager reach the same rating from the evidence?
• Are experienced employees being penalized because their growth is subtler?
• Are junior employees being over-rewarded simply because their learning is more visible?

Calibration should focus on the evidence, not on defending the initial score.

Bias Risks and Safeguards

Growth velocity is vulnerable to bias if left undefined.

Common risks include:

Appendix: Job-Family Customization

It is important to be honest that the current rubric is generic. It can apply broadly across functions, but some organizations may want job-family examples. Job-family customization means keeping the scoring model the same, but changing the examples, evidence expectations, and language so the rubric makes sense for a specific type of work. Do not customize the whole rubric from scratch for each job family. That creates inconsistency. I’d customize only the parts that help managers recognize evidence.

For example, this is what growth can look like for each job family:

Engineering

• Better design documents.
• More reliable delivery.
• Stronger architecture tradeoffs.
• Reduced operational risk.
• Better code review judgment.
• Reusable technical patterns.

Security

• Earlier risk detection.
• Better threat modelling.
• Stronger control design.
• Improved stakeholder influence.
• Better incident learning transfer.
• More scalable security patterns.

Product

• Better prioritization.
• Stronger customer insight.
• Improved discovery quality.
• Better stakeholder alignment.
• Clearer tradeoff decisions.

Operations

• Improved process reliability.
• Better escalation judgment.
• Reduced repeat incidents.
• Stronger documentation.
• More scalable workflows.

The core scoring model can stay the same while examples are adapted to each function.

Example: same dimension, different job families

Take Judgment Improvement.

The generic senior/staff criterion is:

Makes better tradeoffs earlier, with fewer reversals, escalations, or preventable surprises.

That is useful, but still abstract. Job-family customization makes it concrete.

Example: customized evidence artifacts

For Learning Transfer, the generic senior/staff criterion is:

Converts lessons into reusable practices, standards, patterns, decision records, coaching, or team improvements.

That could become:

Again, the scoring model stays the same. The examples become function-specific.