The Metrics That Actually Matter for Product Managers

I've seen teams drown in dashboards. 50 metrics, beautiful charts, weekly reviews — and nobody can tell you if the product is actually doing well.

Metrics aren't the goal. Understanding your product is. Here's what I actually track.

The Problem With Most Metrics

Most metrics fall into two traps:

Vanity metrics: Numbers that go up but don't mean anything. Page views. Registered users. App downloads. They feel good but don't tell you if people are getting value.

Lagging metrics: Revenue. Churn rate. NPS. By the time these move, it's too late to react. You need leading indicators.

What I Actually Track

Depends on the product, but here's my core set:

1. Active Usage (not registrations)

I don't care how many people signed up. I care how many people used the product this week.

• DAU/WAU/MAU — Pick what makes sense for your product
• Engagement depth — Are they using core features or just logging in?
• Usage frequency — Daily users are worth 10x monthly users

For a payments platform I worked on: 10K registrations meant nothing. 2K active merchants processing transactions meant everything.

2. Retention Cohorts

Not just "what's our retention rate" but "how does retention change over time by cohort?"

Plot it:

• Week 0: 100% (they just signed up)
• Week 1: 40% (expected drop)
• Week 4: 25%
• Week 8: 20%

Is that curve flattening or still dropping? If it flattens, you have a core user base. If it keeps dropping, you have a leaky bucket.

I check this monthly. Changes in the curve tell you more than any single number.

3. Activation Rate

What percentage of new users reach the "aha moment"?

You need to define what activation means for your product:

• For a SaaS tool: created their first project
• For an e-commerce platform: made their first purchase
• For a payments product: processed their first transaction

If activation is low, fix that before anything else. Acquisition doesn't matter if people don't stick.

4. Feature Adoption

Which features are actually being used?

I've killed features with 2% adoption. I've doubled down on features with 60% adoption. Usage data should drive roadmap decisions.

Check:

• What % of active users use feature X?
• What's the trend? Growing or shrinking?
• Do power users use it or everyone?

5. Customer Feedback Loop

Not a single NPS number. A continuous stream:

• Support tickets (categorized by theme)
• User interviews (at least 2-3/month)
• Churn reasons (actually ask people why they left)
• Feature requests (weighted by user value)

I review support ticket themes weekly. Spikes tell you something's wrong before metrics do.

Metrics I Stopped Tracking

Page views / session time — Unless you're a content business, these don't matter. Someone could spend 30 minutes frustrated trying to do something simple.

Story points / velocity — Team capacity metric, not product success metric. Delivering 100 story points of the wrong thing is worse than 20 points of the right thing.

Total revenue — Too lagging. By the time revenue drops, the damage happened months ago. Track leading indicators instead.

The Dashboard I Actually Use

One page. Five numbers:

1. Active users (this week vs last week)
2. Activation rate (this month)
3. Retention rate (week 4 cohort)
4. Top 3 feature adoption rates
5. Support ticket volume by category

Plus one qualitative section: notable feedback from this week.

Takes 10 minutes to review. Tells me everything I need to know about whether the product is healthy.

When Metrics Lie

Numbers don't lie, but they can mislead.

Example: Activation rate went up 15%. Celebration? Turns out we lowered the bar for what counted as "activated." Users weren't actually getting more value.

Example: DAU spiked after a press mention. Looked great. But retention for that cohort was terrible. Vanity traffic, not real growth.

Always ask: "What could make this number move without actual improvement?"

The Real Metric

Behind all the numbers: Are users getting value?

If yes, other metrics will follow. If no, no amount of growth hacking will save you. I've seen products with great metrics die because the core value wasn't there. I've seen products with terrible metrics survive because users genuinely loved them.

Track the numbers. But don't mistake the map for the territory.