Introduction
Many founders measure activity. Few measure progress.
A dashboard with fifty numbers usually means nobody knows which five matter. Metrics should help decisions, not create dashboards. Focus creates better decisions, and most teams have far more metrics than focus.
This framework helps you decide which metrics earn their place, which ones do not, and how to keep the dashboard honest as the company grows.
What it solves
What the right metrics deliver
-
Faster, more honest decisions
A small set of well-chosen metrics surfaces decisions weeks earlier than a sprawling dashboard. The team sees the signal because the noise is gone.
-
A shared definition of progress
When every team agrees what 'better' means, planning gets shorter. When they do not, every meeting relitigates first principles.
-
Early warning before problems become incidents
Good metrics show stress before it shows up in revenue or churn. A team without those metrics responds to consequences; a team with them responds to causes.
-
Defensible storytelling for investors and customers
Numbers tied to real customer value land in fundraising and sales. Vanity metrics land flat once the audience is sophisticated.
-
A culture of evidence
Teams that decide with numbers improve faster than teams that decide with opinion. Metrics are how that culture is built, but only if the metrics are the right ones.
What it does not solve
What metrics will not fix
-
An unclear strategy
Metrics describe what is happening; they do not decide what should happen. A startup without a strategy will measure the wrong things confidently.
-
Bad incentives
Whatever you measure becomes what gets done. Measuring volume produces volume; measuring quality produces quality. Pick carefully, because behaviour follows.
-
Customer empathy
Metrics are downstream of conversations. A team that only reads dashboards loses the texture that explains why the numbers move.
-
Slow product decisions
If decisions take weeks, more metrics make them slower, not faster. Decision pace is upstream of measurement quality.
-
A weak product
No metric improves a product that does not yet solve a real problem. Measurement reveals the gap; it does not close it.
Decision tree
Six questions for every metric
Run every candidate metric through these questions. The 'no's are where dashboards quietly turn into wallpaper.
- Question 01
Does the metric influence decisions?
No → Stop tracking it. A metric that does not change behaviour is a number on a screen.Yes → Confirm one decision it changed in the last 30 days. If there is none, the metric is decorative. - Question 02
Does the metric reflect customer value?
No → Be careful. Operational metrics matter, but they are not progress. A startup that improves only operational numbers is moving sideways.Yes → Confirm the customer feels the difference when the metric moves. - Question 03
Can the metric be influenced?
No → Replace it. Metrics outside your control teach helplessness.Yes → Confirm there is at least one action the team can take this week that would move it. - Question 04
Would the business improve if this metric improved?
No → Drop it. A metric whose improvement does not matter is one whose tracking does not either.Yes → Quantify the relationship. 'Probably' is not measurement; it is opinion. - Question 05
Is the metric leading or lagging?
No → Balance them. Too many lagging metrics describe history; too many leading metrics produce speculative dashboards.Yes → Confirm you have at least one leading metric per lagging one. Otherwise you can only react, never anticipate. - Question 06
Would you still track it if nobody asked?
No → Stop tracking it. Metrics maintained for audiences rather than decisions decay quickly.Yes → Confirm the metric earns its place inside the team, not just inside the board deck.
Common mistakes
Five common mistakes founders make
- 01
Vanity metrics
Total signups, total downloads, total page views, large numbers that move in only one direction. They feel like progress but tell you nothing about whether the business is improving. Replace with engagement, retention, conversion or revenue metrics that customers actually drive.
- 02
Tracking everything
Dashboards full of numbers train the team to ignore them. Fewer, better metrics get read; bigger dashboards become decoration. If the team cannot list the top five from memory, the dashboard has too many numbers.
- 03
Reporting without action
Metrics that get reviewed in meetings but never trigger a decision are theatre. If a metric has produced no decisions in three months, either it is the wrong metric or the team is not using it. Both have the same fix: remove it.
- 04
Optimising the wrong number
Hitting a target by gaming a definition is a common failure mode. A 'monthly active user' that counts loading the homepage is not a customer using the product. Define metrics carefully; check what they actually measure quarterly.
- 05
Ignoring customer behaviour
Dashboards summarise; conversations explain. Founders who only read numbers lose the context that turns measurement into decisions. Pair every key metric with a regular customer conversation.
Alternatives
Patterns for a metrics system that works
Four patterns that keep measurement supporting decisions rather than replacing them.
-
The five-metric dashboard
Pick five metrics, usually one for growth, one for retention, one for engagement, one for revenue, one for cost, and put them on a single page. Anything more goes into a secondary dashboard reviewed monthly, not weekly.
-
One North Star metric
Choose a single number that summarises whether the company is succeeding. Everything else supports it. Forces alignment across teams and stops the metrics zoo from growing.
-
Leading + lagging pairs
For every important lagging metric (revenue, churn) keep at least one leading metric (signups, activation, NPS). The lagging metric tells you what happened; the leading one tells you what is about to.
-
Quarterly metric review
Once a quarter, audit every metric. Drop the ones that produced no decisions, add the ones that should have. Treat the dashboard as a living document rather than a permanent installation.
Ronald's rule of thumb
If a metric never changes your decisions, stop tracking it.
Every metric has a cost: the time to instrument it, the cognitive load of reading it, the temptation to optimise it. Metrics that do not change decisions still consume all three. Cutting them is one of the highest-leverage acts in a startup operating system, and one of the least frequently performed.
, Ronald · YourStartup.Expert
Summary
Summary
A useful metrics system answers two questions: is the company improving in a way customers feel, and what should change next week? The six questions above filter every candidate metric down to the ones that answer those questions. Anything else belongs on a secondary dashboard or, more often, in the bin.
Most startups would benefit from measuring less. A small set of well-chosen metrics that the whole team knows by heart produces sharper decisions than a sprawling dashboard nobody can keep current. Discipline in measurement compounds the same way discipline in scope does, and is just as rarely taught.
Common questions
Startup metrics, answered.
The questions founders ask before they design a dashboard.
- What metrics matter most for a startup?
- A small set: one for growth (new customers or users), one for retention (week 4 or month 1 retention), one for engagement (the action customers take when the product is working), one for revenue (MRR, gross margin or LTV) and one for cost (CAC, burn or unit economics). Most early-stage startups should pick five metrics they can recite from memory and ignore the rest until the next quarterly review. Bigger dashboards correlate weakly with better decisions.
- What is a vanity metric?
- A metric that grows reliably without telling you whether the business is improving. Cumulative signups, total downloads, total page views, all rise over time and rarely fall. They feel like progress, but they do not change decisions, do not reflect customer value and do not predict the future. Replace vanity metrics with engagement, retention, conversion or revenue metrics that move in both directions.
- How many metrics should I track?
- Few enough that the whole team can recite them. For most early-stage startups, five top-level metrics on a primary dashboard and another fifteen to twenty on secondary dashboards is plenty. If the dashboard requires scrolling, it has too many numbers. Add metrics deliberately; remove them quarterly.
- What is the difference between leading and lagging metrics?
- Lagging metrics describe what already happened (revenue, churn, NPS). Leading metrics predict what is about to happen (signups, activation rate, support volume). Lagging metrics tell you whether the business is succeeding; leading metrics give you time to react. A healthy dashboard pairs each major lagging metric with at least one leading metric that can be influenced this week.
- When should I review and change my metrics?
- At least once a quarter. The first version of any metric system gets things wrong; quarterly review catches that. Drop the metrics that have produced no decisions, replace them with the ones that should have. Treat metrics as a living system: the goal is to inform decisions, not to maintain a permanent installation.