Product Metrics That Actually Matter (And Ones That Don't)

If you look at a dashboard with fifty different line graphs on it every morning, you aren't being data-driven. You are just looking at a screensaver.

Most product metrics are completely useless. They are designed to make executives feel good in board meetings, not to help product teams make decisions. If a metric goes up, and you don't immediately know what action to take, or if a metric goes down, and you don't change your behavior, then you are tracking the wrong metric.

It is just numbers for the sake of numbers.

Here is how you cut through the noise and find the metrics that actually govern the physics of your product.

The Danger of Vanity Metrics

A vanity metric is a number that always goes up and to the right, regardless of whether your product is actually healthy.

Total Registered Users is the ultimate vanity metric. The only way it ever goes down is if someone explicitly deletes their account. You could have a product where 99% of people churn on day one, but your "Total Users" graph will still look like a rocket ship.

Don't celebrate Total Users. Celebrate Active Users. And even then, be careful how you define "Active."

If "Active" just means they opened the app, that's weak. If I open the Uber app but don't book a ride, I shouldn't count as an active participant in their ecosystem for that day. At Vibo, we realized early on that just logging in meant nothing. The only metric that mattered was whether they completed the core loop.

The North Star Metric

You need one number that aligns the entire company. One number that, if it goes up, means the platform is generally delivering value to both the business and the customer.

• For Spotify, it's Time Spent Listening.
• For Airbnb, it's Nights Booked.
• For WhatsApp, it's Messages Sent.

The North Star shouldn't be Revenue. Revenue is a lagging indicator. It tells you what happened in the past. If you focus exclusively on revenue, you will start optimizing for extraction rather than value creation. You'll put ads everywhere, raise prices, and destroy the core user experience just to hit a quarterly target. Your users will hate you, and eventually, the revenue will collapse.

Focus on the metric that precedes revenue. Focus on the metric that indicates value is being exchanged.

Leading vs. Lagging Indicators

This is the most misunderstood concept in product analytics.

Lagging indicators are the output. They tell you the result of actions taken months ago. Churn rate and Revenue are lagging indicators. You can't directly act on them today. By the time someone cancels their subscription, the damage was done three weeks ago when the app crashed during their presentation.

Leading indicators are predictive. They measure the behaviors that lead to the output.

You can't "fix" churn directly. But you can fix the leading indicators of churn.

• Are users taking more than three seconds to load the dashboard?
• Has a user who usually logs in daily not logged in for three days?
• Have support tickets related to a specific feature spiked?

Those are leading indicators. You can act on them today. If you fix the leading indicators, the lagging indicators fix themselves.

The Retention Curve is Everything

If I could only look at one chart for the rest of my career, it would be the retention curve.

Acquisition metrics are mostly marketing physics. Retention metrics are product physics.

If your day-30 retention is effectively zero, you have a leaky bucket. Stop spending money on acquisition. Stop building new features. Fix the core loop. You don't have product-market fit.

A healthy retention curve drops initially (everyone drops initially when tourists try the app and leave), but then it must asymptote. It must flatten out parallel to the x-axis. If it eventually hits zero, you do not have a sustainable business.

Counter Metrics

Every metric you try to optimize will create negative externalities. The universe demands balance.

If your goal is to increase the conversion rate of a sign-up form, the easiest way to do that is to remove the email verification step. Conversions will skyrocket. But next week, your system will be flooded with spam accounts and your deliverability rate will tank.

You must always pair your primary metric with a counter metric—a metric that ensures you aren't doing something incredibly stupid to hit your goal.

If you are optimizing for speed of delivery (Cycle Time), your counter metric must be Bug Rate. If Cycle Time goes down but Bug Rate goes up, you didn't become more efficient; you just got sloppy.

Measure the action, but always measure the consequence.

---

FAQ

What should an early-stage startup measure?

Before Product-Market Fit, you should measure almost nothing quantitative. If you only have 50 users, your Google Analytics dashboard is meaningless noise. One power user going on vacation will ruin your MAU chart. Focus on qualitative feedback: Net Promoter Score, direct user interviews, and whether they get visibly angry when you threaten to take the product away.

How do I define "Active" for Daily Active Users (DAU)?

Define it as the user completing the core action of your product. For a messaging app, it's sending a message. For a photo app, it's uploading a photo. Logging in is not an action; it is a prerequisite for an action.

We have a data team. Shouldn't they handle the metrics?

The data team owns the infrastructure and the accuracy of the data. You own the interpretation and the resulting product decisions. If you outsource the "why" to the data team, you are abdicating your responsibility as a Product Manager.