The Right KPIs for Product Managers (and Why Most Teams Track the Wrong Ones)

Most product metrics dashboards are anxiety machines.

They show you what happened — pageviews, clicks, feature usage, DAUs — but they can't tell you whether what happened matters. A spike in DAU could mean you shipped something great. It could mean a holiday weekend in a market you weren't targeting. It could mean your email onboarding is pulling in low-intent users who'll churn by Friday.

The problem isn't having metrics. It's having metrics that aren't connected to the outcomes you actually care about.

Here's how to pick product manager KPIs that drive decisions, not just reports.

The Three-Level Metric Stack

Every product metric sits at one of three levels:

Business outcomes — What the company cares about. Revenue, retention, NPS, churn rate. These are the metrics your CEO tracks. They're real and important, but they're lagging indicators — by the time they move, the cause is weeks or months in the past.

Product outcomes — What user behaviour you're trying to change. Activation rate, time-to-first-value, feature adoption, session depth. These are the metrics a PM owns. They're leading indicators for business outcomes if you've chosen them correctly.

Activity metrics — What you shipped. Features launched, bugs closed, sprints completed, experiments run. These are not KPIs — they're output metrics. Measuring them tells you whether your team is busy. It doesn't tell you whether they're building the right things.

Most product manager KPIs live in the activity layer and get mistaken for outcome metrics. "We shipped 14 features this quarter" is not a product KPI. "Feature adoption rate for the new dashboard increased from 12% to 31%" is.

Choosing the Right Product Metrics

The right product metrics for your team depend on your product's growth stage.

Pre-product-market-fit

When you're still searching for PMF, retention is the only metric that matters.

Specifically: do users come back, unprompted, and do what the product is designed for? Not "did they log in" — but "did they complete the core action?"

If retention is flat or declining across all cohorts, adding growth metrics on top is a waste of attention. Fix the bucket before you fill it.

The metric: Week 2 / Week 4 retention rate for activated users. If it's not improving across cohorts, the product isn't working yet.

Post-PMF, pre-scale

Once you have evidence that some users find the product genuinely valuable, your metrics expand:

Activation rate — The percentage of new signups who reach a defined activation threshold within a fixed window (e.g., "generated a PRD within 7 days"). This is where most growth leaks.

Time-to-value — How long it takes a new user to experience the core value prop. For most B2B products, this is the single most improvable metric in the first year.

Feature adoption depth — Not "X users used feature Y" but "what percentage of activated users discovered and used feature Y in month 1?" Depth matters more than breadth.

At scale

When the core loop works and you're optimizing, expansion metrics matter:

Net Revenue Retention (NRR) — The percentage of revenue retained from existing customers after accounting for upgrades, downgrades, and churn. NRR above 100% means your existing customer base is growing without acquiring new customers.

PQL-to-paid conversion — For PLG products: what percentage of product-qualified leads convert to paying? This connects your product team's work directly to revenue.

The North Star Metric

Every product should have one North Star Metric (NSM) — a single number that best represents the value you deliver to users.

The NSM is not a business metric (not revenue). It's not an activity metric (not features shipped). It's the measure of how much value users are getting.

For Spotify: minutes of music listened to per user.

For Airbnb: nights booked.

For PMRead: PRDs generated from customer evidence.

The NSM is useful because it aligns the team. If your engineering, design, and product teams all know the one number that represents whether the product is working, every feature decision can be evaluated against it.

Two tests for a good NSM:

1. 1If this number goes up and everything else stays the same, are we winning? If yes, it's a real NSM. If you can imagine the number going up while the product is failing, it's the wrong metric.
2. 2Can every team member tell you the current number? If it requires a dashboard to answer, it's not simple enough to be a north star.

KPIs for Product Managers by Role

Early-stage PM (0–1 product): Retention and activation. Everything else is noise until these are healthy.

Growth PM: Funnel conversion rates, acquisition quality (do paid users behave like organic users?), virality coefficient if relevant.

Platform PM: API reliability and adoption, developer experience metrics (time-to-first-API-call), dependency health.

Data/ML PM: Model accuracy against business outcome metrics, prediction latency, coverage rate (what percentage of cases does the model apply to?).

Enterprise PM: Time-to-value for new accounts, support ticket volume by feature area, renewal rate, NPS score delta year-over-year.

The Metric Hierarchy in Practice

A common mistake is tracking too many metrics in parallel. If your team has 15 KPIs, you have no KPIs — you have 15 things that are theoretically important, none of which will generate a real decision when they move.

A working metric hierarchy looks like this:

• One North Star Metric — owned by the full product team
• Two or three leading indicators — the inputs that drive the NSM (e.g., "activation rate" and "week-2 retention")
• One guardrail metric per team — the metric you watch to confirm you're not causing harm while improving the primary

Everything else goes into a secondary dashboard that you check when something looks wrong — not as a standing agenda item in every sprint review.

What Bad Metric Selection Looks Like

Vanity metrics: DAU, total signups, total pageviews. These go up when you run ads and down when you turn them off. They don't tell you whether the product is working.

Output metrics dressed as outcomes: "We shipped 3 features." That's not a KPI — it's a task list.

Metrics you can't influence: A solo product team tracking overall company churn when most churn is sales-led is measuring something they can't change.

Too many primary metrics: "Our team tracks 12 KPIs." No team can optimise 12 things at once. Pick the 2 that matter most for the next 90 days and let the others be monitoring metrics.

The test: if your KPI moves significantly this week, would you know what to do differently? If not, it's not a useful KPI — it's a dashboard decoration.

Link your metrics to your product strategy in your North Star Metric template to make the connection explicit.