Not all metrics are created equal. Most dashboards measure activity, not impact. Here's how to tell the difference — and how to build a metrics framework that actually guides product decisions.
I've sat in quarterly business reviews where teams proudly reported 40% growth in feature usage, 25% increase in time-on-product, and a 15-point NPS improvement. Then the CFO asked a simple question: "Is revenue growing?" Silence. Revenue was flat. The product was generating more engagement, not more value. The team had been optimizing for the wrong things and calling it success.
This is the metric selection problem. Every metric you measure is a choice about what matters. Choose wrongly, and you build a team that is diligent, organized, and working hard at the wrong goals. Choose well, and the metrics themselves create alignment — because everyone can see what success looks like and whether you're moving toward it.
The Metric Hierarchy
The most useful mental model for metrics is a hierarchy: each level should causally connect to the one above it. At the top is your ultimate business outcome — revenue, margin, market share, or whatever the business fundamentally optimizes for. Below that is the user outcome — the change in user behavior that drives the business outcome. Below that are the leading indicators — the product behaviors that predict the user outcome. At the bottom are the operational metrics — the things you track to ensure the product is functioning correctly.
When metrics at different levels of this hierarchy are misaligned — when leading indicators are improving but user outcomes aren't, or when user outcomes are improving but business outcomes aren't — that's a signal that either your metrics are wrong or your theory of how the product creates value is wrong. Both are important to know.
The Metrics Most PMs Get Wrong
DAU and MAU
Daily and monthly active users are the most cited product metrics and among the least useful in isolation. They tell you how many people showed up — not whether they got value, whether they'll come back, or whether they're paying customers. A product with 1 million MAU and 5% retention is a leaky bucket. A product with 100,000 MAU and 80% retention is a compounding machine. MAU alone doesn't tell you which one you have.
Session length
Longer sessions are better. Usually. Except when long sessions mean users can't find what they're looking for. Or when your product is a utility where speed of task completion is the value proposition. Time-on-product as a success metric is an Instagram-era assumption that doesn't apply to most products. Ask: for what type of product is longer session time actually better? Only track it if the answer is clearly yours.
Feature adoption rate
X% of users have tried feature Y. This tells you about awareness and initial use — it says nothing about whether the feature is creating value. A feature that 30% of users try once and never use again has 30% adoption. A feature that 15% of users use weekly and credit for their continued subscription has 15% adoption. Which one matters more? Track retention and frequency of use, not one-time adoption.
Building a Metric Framework That Works
Start by answering: what is the one thing that has to be true for this product to be valuable? For a messaging app, it might be: messages get delivered and read quickly. For a task manager, it might be: tasks get completed. For a fintech product, it might be: users make better financial decisions. Whatever that core thing is, build your primary metric around it.
Then build guardrails: metrics that tell you when something is going wrong even if your primary metric looks fine. Revenue is great, but if NPS is collapsing, something is wrong that will show up later. Activation is improving, but if support tickets are spiking, something needs attention. Guardrail metrics catch what your primary metrics miss.
Finally, assign ownership. Every metric should have a human being who is accountable for it. Not "the team" — a specific person whose decisions primarily drive that metric. When metrics don't have owners, they don't move on purpose. They move because of random variation or external factors, and nobody learns anything from them.
Working on your metrics framework?
The right metrics framework is specific to your product's business model and growth stage. I've helped teams identify the metrics that actually drive their business and build dashboards around them. Let's build yours.
Work with Yogesh →