Google Analytics shows 50,000 pageviews last month.
Your boss is excited. Marketing declares victory.
Meanwhile, MRR stayed flat. Churn went up. Trial-to-paid conversion dropped.
Something’s broken. And it’s not the traffic.
The data delusion
Tech companies drown in data while starving for insight.
Every tool promises “actionable analytics.” Dashboard after dashboard. Metrics everywhere.
Pageviews. Bounce rate. Time on site. Session duration. Scroll depth. Heat maps. Click maps. Funnel drop-offs.
None of it tells you why revenue isn’t growing.
Here’s the uncomfortable truth: most analytics setups measure activity, not outcomes.
Activity metrics feel productive. They go up. They generate reports. They fill slide decks.
But activity doesn’t equal progress.
A SaaS company can have amazing traffic, perfect engagement metrics, and still fail because they’re not tracking what actually drives the business.
The metrics that actually matter for tech products
Forget vanity metrics. Focus on revenue-connected data.
For B2B SaaS, these are the fundamentals:
Monthly Recurring Revenue (MRR). The only number that truly matters. Everything else supports understanding how to grow this.
Customer Acquisition Cost (CAC). What does it actually cost to acquire a paying customer? Include everything: ads, tools, salaries, agency fees.
Lifetime Value (LTV). How much revenue does a customer generate before churning? If LTV is lower than CAC, you don’t have a business model.
Churn rate. How many customers leave each month? High churn means product-market fit issues, not marketing issues.
Activation rate. What percentage of signups actually complete meaningful actions in your product? This reveals if your onboarding works.
Time to value. How long does it take a new user to experience the core benefit? Shorter is better.
Feature adoption. Which features correlate with retention? Double down on those in marketing.
These metrics tell a story about business health. Pageviews don’t.
Google Analytics is not enough
GA4 is powerful. It’s also overwhelming and incomplete for tech products.
It tracks website behavior. But for SaaS, the website is just the entry point. The real product experience happens after login.
What GA4 doesn’t tell you:
How users actually interact with your product. Which features they use. Where they get stuck. What correlates with upgrade decisions.
For that, you need product analytics.
Mixpanel, Amplitude, or PostHog track user behavior inside your product. Events. Funnels. Cohorts. Retention curves.
This is where you discover:
Users who complete onboarding within 24 hours have 3x higher retention. Users who invite teammates convert to paid at 5x the rate. Power users spend 80% of their time in two specific features.
These insights shape product roadmap and marketing messaging.
Without product analytics, you’re flying blind.
Event tracking is where insight lives
Generic pageview tracking is lazy.
Event tracking is strategic.
Every meaningful action in your product should be an event:
- User completes signup
- User creates first project
- User invites team member
- User hits usage limit
- User views pricing page
- User starts trial
- User integrates with external tool
- User completes core workflow
Track these events. Analyze patterns.
You’ll discover: users who create a project within their first session have 60% higher conversion rates.
That insight changes everything. Now you optimize onboarding to push users toward creating that first project immediately.
One data point. Massive impact.
That’s the difference between tracking and measuring.
Cohort analysis reveals truth
Aggregate metrics lie.
“Our conversion rate is 5%!” Sounds good. But what if:
Cohort 1 (users from paid ads) converts at 2%. Cohort 2 (users from organic content) converts at 12%.
Aggregated, it’s 5%. But the story is completely different.
Cohort analysis segments users by when they signed up or what acquisition channel brought them.
This reveals:
Which marketing channels actually work. How product changes impact new vs existing users. Whether recent cohorts perform better than old ones (sign of improvement).
Without cohorts, you’re averaging away your most important insights.
Attribution is broken (but you still need it)
Multi-touch attribution is mostly fantasy.
The idea: track every touchpoint in a customer’s journey and assign value to each.
Reality: people browse in incognito mode. Switch devices. Research on mobile, convert on desktop. Use ad blockers.
Perfect attribution is impossible.
But you still need to understand influence:
First-touch attribution: what initially brought them in? Last-touch attribution: what finally converted them?
Neither tells the full story. Both provide context.
Better approach: ask customers how they found you. Add a field in your signup flow. Send a survey post-purchase.
Qualitative data fills gaps that analytics miss.
Reporting is communication, not documentation
Most reports are data dumps.
37 slides. Every metric imaginable. No story. No insights. No recommended actions.
These reports get ignored.
Effective reporting answers specific questions:
What happened? (the data) Why did it happen? (the analysis) What should we do about it? (the recommendation)
A good report is 3-5 slides with clear narrative:
MRR grew 8% this month. Growth came primarily from expansion revenue, not new customers. Enterprise segment showed strongest performance. Recommendation: increase investment in enterprise marketing.
That’s actionable. That drives decisions.
Everything else is noise.
Real-time dashboards for teams
Reports are for stakeholders. Dashboards are for operators.
Every team member should have access to live data relevant to their role:
Marketing sees: traffic sources, conversion rates, CAC by channel, content performance.
Sales sees: pipeline value, deal velocity, win rate, average contract value.
Product sees: feature usage, activation rates, retention curves, bug reports.
Support sees: ticket volume, response time, CSAT scores, common issues.
When data is visible and current, decisions improve. Teams self-correct. Problems surface faster.
Transparency beats control.
Custom events for specific insights
Off-the-shelf analytics track generic actions.
Custom events track what’s unique to your product.
If you’re building a code editor, track:
- Lines of code written
- Debugging sessions started
- Keyboard shortcuts used
- Extensions installed
If you’re building project management software, track:
- Tasks created vs completed
- Team collaboration frequency
- Project templates used
- File uploads and sharing
These custom metrics reveal user behavior patterns that generic analytics miss.
They inform product development. Marketing messaging. Sales enablement.
They’re also competitive advantages—insights your competitors don’t have.
The danger of too many metrics
Paradox: more data often leads to worse decisions.
When everything is measured, nothing is prioritized.
Teams chase improving 15 different metrics simultaneously. They optimize locally while missing global picture.
Better approach: identify 3-5 North Star metrics.
These are the metrics that best indicate business health and growth.
For an early-stage SaaS: Weekly Active Users, Activation Rate, MRR Growth might be your North Star.
Everything else is supporting data. Helpful for diagnosis. But not the primary focus.
Clarity beats comprehensiveness.
Instrumentation requires planning
Analytics isn’t something you bolt on after launch.
It requires intentional instrumentation from day one.
Before building features, define:
What success looks like. What behaviors indicate value. What metrics will measure progress.
Then instrument tracking for those specific things.
This requires collaboration between product, engineering, and marketing.
Engineers implement tracking. Product defines what to track. Marketing interprets results.
Without this alignment, you get inconsistent data, missing events, and analysis paralysis.
Privacy and compliance matter
GDPR. CCPA. Cookie consent. Data retention policies.
Analytics can’t ignore legal requirements.
Users have rights: to know what’s tracked, to opt out, to request data deletion.
Compliance isn’t optional. And users increasingly care about privacy.
Best practices:
Be transparent about what you track. Allow opt-out. Don’t store PII unnecessarily. Use privacy-respecting tools when possible.
Some companies now use server-side analytics to reduce client-side tracking. Others choose privacy-focused alternatives like Plausible or Fathom.
The goal: get insights without exploiting users.
That balance is increasingly important.
Testing beats guessing
Analytics tells you what’s happening. Testing tells you what to do about it.
A/B testing. Multivariate testing. Feature flags.
Hypothesis: changing CTA color will improve conversion.
Don’t guess. Test it. Let data decide.
Tools like Optimizely, VWO, or LaunchDarkly make this accessible.
But testing requires discipline:
Run tests long enough for statistical significance. Don’t call winners early. Test one variable at a time. Document everything.
The companies that grow fastest are the ones that test relentlessly.
The analytics stack for tech products
You don’t need 47 tools. You need the right ones:
Website analytics: GA4 or Plausible for traffic and acquisition.
Product analytics: Mixpanel, Amplitude, or PostHog for in-app behavior.
Session recording: Hotjar or FullStory to watch actual user sessions.
Error tracking: Sentry or Rollbar to catch technical issues.
Customer data platform: Segment to unify data across tools.
Business intelligence: Metabase or Looker for custom dashboards.
Spreadsheets: Still the most versatile analysis tool. Don’t underestimate Excel/Sheets.
Pick tools that integrate well. Avoid redundancy. Focus on actionable insights, not data collection for its own sake.
Learning from outliers
Averages hide extremes.
Average customer pays $50/month. Sounds fine.
But 90% pay $10. 10% pay $500.
That changes strategy completely. You’re not targeting the average. You’re targeting the high-value segment.
Same with usage patterns:
Average user logs in 3x per week. But power users log in daily and generate 80% of your referrals.
Outliers aren’t noise. They’re often your most important signal.
Study power users. Understand what makes them different. Figure out how to create more of them.
That’s how you scale.
Automation reduces busywork
Manually pulling reports every week is a waste of time.
Automate repetitive tasks:
Scheduled reports that email stakeholders. Alerts when metrics hit thresholds. Dashboards that update in real-time.
This frees bandwidth for actual analysis.
The goal isn’t eliminating human oversight. It’s eliminating human drudgery.
Let machines handle routine monitoring. Let humans focus on interpretation and strategy.
From data to decisions
Analytics only matters if it changes behavior.
A report that generates no action is worthless.
Every analysis should end with: “Based on this data, we should…”
Increase budget on X channel. Redesign Y feature. Deprecate Z workflow. Test A hypothesis.
Data informs decisions. Decisions drive outcomes.
If your analytics setup isn’t directly influencing what your team builds, markets, or sells, it’s not working.
Fix the connection between insight and action.
That’s where analytics becomes valuable.