Your product analytics dashboard tracks 47 different metrics. Page views, session duration, feature clicks, API calls, user segments, cohort comparisons. The dashboard looks sophisticated. Stakeholders are impressed by the depth of instrumentation.
But when it's time to make a product decision, nobody knows which metrics actually matter. Is a 15% increase in session duration good or bad? Should you optimize for feature clicks or feature completion? Does higher API usage indicate engagement or inefficiency?
Tracking everything feels comprehensive, but it's actually paralyzing. The more metrics you track without clear purpose, the harder it becomes to identify signal from noise.
After helping six B2B companies overhaul their analytics strategy, I've learned the hard way: fewer metrics with clear purpose beats more metrics with vague intent. Every time.
Here's how to identify the 3-5 metrics that actually drive your business and eliminate the rest.
The Vanity Metric Test
A vanity metric is any number that makes you feel good but doesn't inform decisions. They're easy to track, often trending upward, and completely useless for product strategy.
Run every metric through this three-question test:
Question 1: If this metric goes up, do we know what action caused it?
If the answer is "not really," it's a vanity metric. Page views might be up because you launched a new feature, or because your most engaged users are visiting more, or because bot traffic increased. Without attribution to specific actions, you can't learn from the change.
Question 2: If this metric goes down, do we know what action to take?
If the answer is "we'd have to investigate," it's a vanity metric. Session duration dropping might mean users are finding what they need faster (good) or getting frustrated and leaving (bad). The metric alone doesn't tell you whether to celebrate or panic.
Question 3: Would we make a different decision if this metric changed significantly?
If the answer is "probably not," it's a vanity metric. Many teams track metrics they'll never act on. Total registered users might be an interesting number, but if you're focused on paid conversion, free user counts don't change your roadmap.
Vanity metrics survive because they feel like progress. They trend upward over time as you grow. But upward trends without actionable insights waste analytics resources on measurement that doesn't drive improvement.
The Three Types of Metrics That Actually Matter
Every SaaS product needs three types of metrics. Not three dozen. Three types, with 1-2 metrics per type.
Type 1: Activation Metrics (Do users reach value?)
These measure whether users accomplish the core outcome your product promises. Not whether they click around or explore features, but whether they achieve the result they signed up for.
For a CRM: "Created first deal and logged first customer interaction within 7 days"
For a marketing platform: "Launched first campaign and got measurable results within 14 days"
For an analytics tool: "Connected data source and viewed first actionable insight within 24 hours"
Activation metrics are binary: the user either reached the value moment or didn't. This clarity makes them actionable. If activation rates drop, you know onboarding is broken. If they rise, you know you've improved time-to-value.
Type 2: Engagement Metrics (Do users form habits?)
These measure whether users return and build the product into their workflow. Active usage that indicates the product has become indispensable, not just occasionally useful.
The key is measuring qualified engagement, not just any activity.
Don't track: Daily active users (tells you nothing about value)
Track: Users who completed [core action] 3+ times this week (indicates habit formation)
Don't track: Features clicked (tells you nothing about outcomes)
Track: Users who achieved [desired outcome] using the product (indicates value delivery)
Don't track: Session duration (tells you nothing about efficiency)
Track: Time from login to completing [key workflow] (indicates product efficiency)
Engagement metrics reveal whether your product is a nice-to-have or a must-have. Nice-to-haves get used occasionally. Must-haves get used repeatedly in consistent patterns.
Type 3: Retention Metrics (Do users stick around?)
These measure whether users who activated and engaged continue using the product over time. This is the ultimate validation that you're delivering sustained value.
The most useful retention metric isn't a simple percentage. It's cohort-based retention that shows when and why users churn.
Track: "Week 1, Week 4, Week 12, Week 24 retention rates by cohort"
This structure reveals patterns. If Week 1 retention is 90% but Week 4 drops to 60%, your onboarding works but your core value prop doesn't sustain. If retention is high through Week 12 but drops at Week 24, you have a long-term engagement problem.
Segment retention by key user characteristics: plan type, company size, use case, acquisition channel. This shows you which user segments naturally retain and which need intervention.
The North Star Metric Framework
Once you've identified activation, engagement, and retention metrics, choose one North Star Metric that represents product success in a single number.
Your North Star must satisfy three criteria:
Criteria 1: It measures value delivery, not activity
Bad North Star: Monthly active users (tells you nothing about value)
Good North Star: Monthly active users who completed [core workflow] 4+ times (indicates value delivery and habit formation)
Criteria 2: It's leading, not lagging
Bad North Star: Annual recurring revenue (you can't optimize product for revenue directly)
Good North Star: Weekly active teams creating and sharing [key output] (this behavior predicts revenue expansion)
Criteria 3: Every team can impact it
Your North Star shouldn't be owned by one function. Product, marketing, sales, and customer success should all have levers they can pull to move the metric.
Example North Star metrics that work:
- For Slack: Daily active teams sending 2,000+ messages per month
- For Stripe: Weekly active businesses processing payments
- For Notion: Weekly active workspaces with 3+ users collaborating
These metrics indicate value delivery, predict revenue, and give every team a clear target to optimize for.
How to Sunset Metrics Without Causing Panic
Once you've identified your core metrics, you need to stop tracking the vanity metrics. This is harder than it sounds. Teams get attached to metrics they've tracked for months or years.
Step 1: Categorize every current metric
Create three lists:
- Core metrics: The 3-5 that inform key decisions (keep these)
- Diagnostic metrics: Helpful for specific investigations but not for ongoing monitoring (archive these, don't display them in dashboards)
- Vanity metrics: Numbers that don't inform decisions (delete these entirely)
Be ruthless. If you can't articulate how a metric drives a specific decision, it's not a core metric.
Step 2: Communicate the why
Teams resist change when they don't understand the reason. Explain that fewer metrics means clearer focus, faster decisions, and better outcomes.
"We're simplifying our analytics to focus on the metrics that actually drive our business. This means we'll stop tracking [vanity metrics] and focus entirely on [core metrics]. This will help us make faster decisions and improve what actually matters to customers."
Step 3: Run a 30-day test
Don't permanently delete metrics immediately. Hide them from primary dashboards but keep them accessible for 30 days. If nobody asks about them, they weren't actually important.
The Weekly Metrics Review Ritual
Once you've narrowed to your core metrics, establish a weekly ritual for reviewing them as a team.
Every Monday, review the same four questions:
- Which core metric moved significantly this week? (Up or down more than expected)
- What do we think caused the movement? (Hypothesize before diving into data)
- What additional data would confirm or refute our hypothesis? (Diagnostic metrics)
- What should we do differently this week based on what we learned? (Action items)
This ritual transforms metrics from passive reporting to active learning. You're not just tracking numbers. You're using data to improve the product every single week.
When you cut the noise and focus on signal, analytics becomes a competitive advantage instead of a reporting burden.