Every startup dashboard I've seen has too many metrics on it. Founders start tracking everything they can measure — page views, likes, followers, signups, downloads, MAUs, DAUs, CTR, bounce rate, feature usage — because measuring feels productive. In practice, most of it is noise that lets you avoid the two or three uncomfortable numbers that actually matter.
I've watched founders raise money on "growth" that turned out to be free-tier signups no one converted from. I've watched startups celebrate quarter-over-quarter user growth while burning through cash faster than the new users could ever justify. I've watched teams optimize the wrong metric for months and mistake activity for progress.
Here's the honest list of what actually matters — for a startup earlier than Series B — and the metrics that feel important but aren't.
First: understand the difference between metrics and vanity metrics
A real metric answers: "if this number changed, would my business be meaningfully different?"
A vanity metric answers: "does this number make me feel like the business is going well?"
The tell: if a metric only ever goes UP (page views, cumulative signups, total downloads), it's usually vanity. Real metrics can go down. That's the point — they tell you truth about the business, which is sometimes uncomfortable.
Founders who track only vanity metrics run into a specific trap: they optimize for the metric, the metric grows, and the business doesn't. Six months later they discover their fundraising story doesn't survive due diligence.
The 5 metrics that actually matter
Track these five. In this order. If you can only track one, track #1.
1. Weekly active revenue / paid retention
What it is: how much revenue you're earning from customers who are actively using and paying for your product, measured week over week.
For SaaS: Monthly Recurring Revenue (MRR), but broken down into new/expansion/churn/reactivation. For e-commerce: revenue from repeat customers vs first-time buyers. For services: revenue from clients still on retainer this month.
Why it's #1: this single number tells you whether your business is alive. If it's growing, real people are getting real value they'll pay real money for. If it's flat or declining, everything else is theater.
What to look at specifically:
- Total MRR trend (or equivalent) week over week
- The four components: new / expansion / churn / reactivation
- The ratio: for every $1 you're losing to churn, are you gaining $2+ in new + expansion?
When it goes wrong: you're gaining new customers but losing existing ones equally fast. Net growth = zero. This is the trap that kills more startups than any other. New logos feel good on a dashboard; without retention, they're a treadmill.
2. Activation rate (or its equivalent for your business)
What it is: the percentage of new signups (or new customers) who reach the first meaningful moment of value.
For SaaS: users who completed the core action the product exists to help with, within a defined window (e.g., "created and shared their first document within 7 days of signup"). For e-commerce: first-time buyers who complete a second purchase within 60 days. For services: clients who renew after the initial engagement.
Why it matters: activation is a leading indicator of retention. Retained users came from activated users. Activated users came from signups. If your activation rate is broken, no amount of marketing spend will save you — you're pouring users into a bucket with a hole in the bottom.
Benchmarks:
- SaaS median activation: 20-40%
- Best-in-class: 60%+
- Below 15%: something is broken in onboarding or product-market fit
If you don't yet know your activation event, that's your first project this week. The SaaS onboarding fixes post on this site covers how to find and instrument your activation event specifically.
3. Cash runway
What it is: how many months of cash you have left at your current burn rate.
Why it matters: you can be building a great product with growing revenue and still die because you ran out of money before you got to sustainable. Runway is the boundary condition on every other decision you make.
How to calculate:
Cash in bank / Monthly burn = Runway (in months)Where "monthly burn" = monthly costs minus monthly revenue (if revenue > costs, congratulations, you're not burning).
What to actually watch:
- Runway trend (is it getting shorter or longer? By how much per month?)
- Cash-out date (the specific month you run out of money if nothing changes)
- Contingency scenarios (what does runway look like if your top customer churns, if a hire doesn't work out, if a planned round takes 3 more months?)
When it's wrong: many founders track "average monthly burn." That's the wrong metric. What matters is: given current trajectory (not average), when do I run out of money? The trajectory can be much worse than the average.
4. Customer acquisition cost (CAC) vs lifetime value (LTV)
What it is: how much it costs to acquire a customer, compared to how much revenue that customer will produce over their lifetime with you.
Why it matters: this ratio tells you if your business model actually works at scale. If CAC > LTV, you're losing money on every customer — growing faster doesn't help, it hurts.
How to calculate:
- CAC: total sales + marketing spend for a period, divided by new customers acquired in that period. Include EVERYTHING: ad spend, tools, salespeople's salaries, agency fees, referral bonuses.
- LTV: average revenue per customer per month × average customer lifespan (in months) × gross margin percentage.
The rule: LTV should be at least 3x CAC. Below that, the unit economics are marginal. Below 1x, the business is fundamentally broken.
Common mistake: calculating CAC only from ad spend and forgetting the salespeople, marketing tools, agency retainers, and free trial support costs. Real CAC is almost always 2-3x higher than founders think.
5. Time to complete the core user action
What it is: how long it takes a user to do the ONE thing your product is supposed to help them with, measured end-to-end.
Examples:
- For a scheduling tool: time from opening the app to publishing a completed schedule
- For a support tool: time from ticket opened to resolved
- For an e-commerce site: time from landing to completed checkout
- For a data tool: time from data uploaded to first useful insight
Why it matters: every second you shave off is more likely users will actually complete the action. And this metric surfaces friction faster than any survey — if the time is getting slower over releases, you added complexity that isn't earning its weight.
The specific tell: if your product has grown a lot of features but the "time to core action" hasn't improved (or has gotten worse), you're adding functionality that's decreasing user success. That's the pattern behind many startup product deaths — accumulating features, losing focus.
The 25 vanity metrics to stop obsessing over
Not all of these are useless. Some are useful in context. But if you track them WITHOUT the 5 above, you're deceiving yourself.
Traffic vanity (10 metrics): 1. Total page views 2. Total unique visitors (this month, this year, ever) 3. Cumulative signups (of course it goes up over time) 4. Cumulative downloads 5. Bounce rate on a single page (context-free) 6. Time on site (users might be lost, not engaged) 7. Number of countries you have users in (spam bots count) 8. Impressions on your ads 9. Reach on social media 10. Ranking for a keyword (without traffic to prove it matters)
Social vanity (7 metrics): 11. Followers on any social platform 12. Likes on posts 13. Shares of a blog post (if the shares don't drive real traffic) 14. Newsletter subscribers (without open rates) 15. Waitlist size (without conversion when you launch) 16. Discord/Slack member count 17. Press mentions (without measurable impact)
Product vanity (8 metrics): 18. Total number of features shipped 19. Lines of code written 20. Number of releases per week 21. Beta tester count (without activation) 22. App store rating (if you have 30 reviews from friends) 23. NPS from your happiest customers only (survivor bias) 24. Feature adoption of a specific new feature (without connecting to retention) 25. Support ticket volume decreasing (might mean users gave up)
The test: if any of these numbers doubled tomorrow, would your business be meaningfully better? If the answer is "no, I'd feel better but nothing real would change," it's vanity.
How to actually set up a dashboard
Given the 5 metrics + 25 vanity metrics — what should your actual dashboard look like?
One page. Big numbers. No color coding beyond good/bad. Updated weekly.
At the top:
- Runway (specific number of months, and the cash-out date)
- Weekly revenue (this week, last week, delta)
- Activation rate (this week's cohort, previous week's cohort, delta)
Below (secondary):
- CAC vs LTV (current ratio, trend over last 6 months)
- Time to core action (this week's average, previous week's average)
Below that:
- Two or three metrics specific to your product's core loop — e.g., for a marketplace: liquidity (percentage of supply matched to demand); for a content product: content shipped that got engagement; for a service business: capacity utilization.
That's it. No 30-metric dashboards. No overlapping charts. If you can't fit it on one screen, you're tracking too many things.
The metric conversation with your team
Once a week, everyone in the company looks at these numbers together. This has three effects:
1. Alignment: everyone knows what the business is optimizing for. Marketing knows why they exist (CAC + activation). Product knows why they exist (activation + retention). Sales knows why they exist (revenue quality, not just volume).
2. Truth: the whole team sees the honest numbers, including the bad ones. You can't hide from bad activation numbers when they're on the wall every Monday.
3. Decisions: every big decision (should we hire? should we spend on ads? should we build this feature?) can be tested against "will this move one of these metrics?" If it won't, don't do it.
Founders who track many metrics have hazy conversations about "growth" and "engagement." Founders who track 5 have specific conversations about which specific number needs to move and how.
Common metric mistakes
Mistake 1: Averaging over time instead of tracking by cohort.
"Our activation rate is 30%" hides everything. Activation rate for the cohort that signed up in January might be 45% (before you broke something), while February's cohort is 18%. Track by cohort, not by rolling average.
Mistake 2: Vanity in disguise.
"Total revenue this month" sounds real but might be one big enterprise deal masking that all your smaller customers are churning. Break out: revenue by cohort, by plan, by segment.
Mistake 3: Not looking at the metric that scares you.
Every founder has one metric they secretly don't want to check. It's usually churn or CAC or runway. That's the one you have to look at first. The metrics you avoid are the ones that will kill the business.
Mistake 4: Tracking too early.
If you have 50 users total, activation rate is not statistically meaningful yet. Focus on qualitative feedback first (talk to users). Metrics become useful once you have hundreds of users, not tens.
Mistake 5: No connection between metrics and decisions.
If your dashboard shows numbers but no decision you'd make would ever be different based on what it shows, you're not tracking — you're doing performative measurement. Every metric on your dashboard should have a "we'd do X if this hit Y" attached.
A specific example
I worked with a SaaS founder who had a 30-metric dashboard. Monthly active users, daily active users, page views, blog signups, feature usage across 12 features, NPS, support ticket volume, average session length, feature discovery rate — a full page of numbers.
We stripped it down to 5:
- MRR (weekly, broken into new/churn/expansion)
- 7-day activation rate
- Runway
- CAC vs LTV
- Time from signup to first "aha moment" (they defined this as user's first shared document with a collaborator)
Within 8 weeks:
- The team realized activation was actually 22%, not the 45% they'd been assuming
- Churn was hidden inside the "MAU growth" narrative — turned out they were losing customers faster than gaining them by revenue (adding small customers, losing bigger ones)
- CAC calculation revealed they were paying 4x LTV — the business fundamentally didn't work at current pricing
Result: they raised prices 3x for enterprise plans, fixed a broken onboarding step that had been depressing activation, and stopped a $15k/month ad campaign that was acquiring users who never activated.
None of these decisions were visible in the old 30-metric dashboard. They only became visible when the noise got removed.
What to do this week
Delete the metrics you're currently tracking that fail the vanity test. Be brutal. If a metric changing wouldn't change any decision you'd make, remove it.
Define your activation event in one sentence. The specific action a new user takes that predicts they'll stick around. If you don't know, look at your data: what did your retained users do in their first week that churned users didn't?
Calculate your real CAC (including ALL costs) and your real LTV. Do it honestly. If the ratio is below 1x, you have a business model problem, not a marketing problem. If it's below 3x, you have unit economics that need work.
Calculate your runway TODAY. Not from an old spreadsheet. Current cash divided by current burn = current runway. Note the cash-out date on your calendar.
Time yourself doing your product's core action. As a new user. From landing on the homepage to completing the core action. That's your baseline. Every product change you ship should either speed this up or leave it alone — never slow it down.
If you're not sure what your product's core action even IS, that's a bigger problem — the PRD guide on this site covers how to name it precisely. And if you're deciding whether to build product analytics tooling yourself vs subscribing to something like Mixpanel or Amplitude, the build vs buy framework is the prior decision. Product analytics is a classic "buy, don't build" category.
Tracking 5 metrics doesn't mean tracking less carefully. It means tracking more carefully. Every metric on the dashboard has to earn its spot by being connected to a real decision. Everything else is noise that helps you avoid the truth.
---
If you want an outside opinion on which 5 metrics ACTUALLY matter for your specific business, reach out via the contact page with a paragraph about what you're building and who you're serving. I'll spend 30 minutes on it and tell you what I'd track — and what I'd stop tracking — if it were mine.