Early-Stage Startup Metrics: Seven Numbers That Matter
Runway, activation, retention, feedback quality, and convert intent.
An early SaaS company can measure hundreds of things and learn nothing from any of them. Page views, total signups, social followers and time on site all move, all feel like progress, and almost none of them tell a founder what to do on Monday.
This post picks seven numbers for a small B2B SaaS selling to small businesses, from the first paying customer to roughly the point where you hire your first non-founder. Each one earns its place because a change in it should change a decision. For each, you get a definition, an example of how to calculate it, and a note on what it is telling you.
The running example is a property inspection app for small real estate agencies: property managers use it on a phone to record routine inspections with photos and send the report to the landlord. All figures below are made up to show the arithmetic.
1. Weekly accounts doing the core job
Definition: the number of customer accounts that completed the product's core job at least once in the last seven days. For the inspection app, that means an inspection report sent to a landlord, not a login.
Example: 34 agencies are signed up. 21 sent at least one report this week. Your number is 21, not 34.
Why it matters: it is the most honest single measure of whether the product is part of someone's work. Logins can be accidental. Sending a report to a landlord is not.
2. Activation rate
Definition: of the accounts that signed up in a given period, the share that completed the core job for the first time within a set window.
Example: 12 agencies signed up in March. 7 sent their first real inspection report within 14 days. Activation is 7 out of 12, about 58 percent.
Why it matters: activation is where most early products leak. If it is low, spending more on marketing mostly buys you more people who sign up and leave. Choose the window to match how often the job happens: inspections happen weekly at a busy agency, so 14 days is fair. A monthly job needs a longer window.
3. Cohort retention
Definition: of the accounts that activated in a given month, the share still doing the core job two and three months later.
Example:
| Activated in | Accounts | Still active at month 2 | Still active at month 3 | |---|---|---|---| | January | 8 | 7 | 6 | | February | 10 | 8 | 8 | | March | 7 | 6 | not yet |
Why it matters: retention is the clearest early signal of product value. What you want to see is the line flattening: the accounts that survive the first couple of months mostly stay. If each cohort keeps sliding towards zero, no amount of acquisition fixes it. Track accounts rather than users, because in B2B one account can have several people.
4. Monthly recurring revenue, and where it came from
Definition: the total monthly subscription revenue from active paying accounts, split into new, expansion, contraction and churned.
Example: you start the month at $2,400. You add $450 from new agencies, $120 from existing agencies adding property managers, lose $80 to a downgrade and $150 to a cancellation. You end at $2,740, and net new revenue for the month is $340.
Why it matters: the total alone hides the story. A month with strong new sales and heavy churn looks similar to a quiet month with none of either, and they need opposite responses. Also note what it costs you to acquire the revenue: if a channel or partner takes a cut, count that. For example, if you convert a beta customer who pays $50 a month and a marketplace takes 15 percent of their first six months, that is $45 in total, which you should subtract when comparing channels.
5. Runway in months
Definition: cash in the bank divided by average net monthly burn (spending minus revenue) over the last three months.
Example: $60,000 in the bank, average spend of $9,000 a month and revenue of $2,700 a month gives a net burn of $6,300. Runway is about 9.5 months.
Why it matters: every other number is a question about direction. Runway is the deadline for answering it. Bootstrapped founders often skip this because there is no investor asking. That makes it more important, not less, because the only warning you will get is the one you build yourself.
6. Customer conversations per week
Definition: the number of real conversations with customers or prospects in your target segment in the last seven days. Calls, visits, detailed feedback exchanges. Not support tickets closed.
Example: two onboarding calls, one visit to an agency, three follow-ups on feedback. Six.
Why it matters: at this stage, the quality of your decisions depends on how close you are to customers. This is the one metric that is entirely in your control and the one that falls first when a team gets busy building. Setting yourself a weekly floor, say five as a rule of thumb, keeps it from quietly dropping to zero.
7. Share of new customers from word of mouth
Definition: of the new paying accounts this month, the share that came from a referral, a recommendation or someone who heard about you from another customer. Find out by asking every new customer how they heard about you.
Example: 4 new agencies this month. 2 said another agency told them. 50 percent.
Why it matters: when customers bring customers, you are becoming known inside your segment, which is one of the strongest signs you are near product-market fit. When the number stays at zero despite happy users, the product may be useful but not remarkable, or your customers may not talk to each other much, which is a segment question.
What to leave off the dashboard
Some numbers are worth knowing occasionally but should not drive weekly decisions early on:
- Total signups ever. Only goes up. Tells you nothing about this week.
- Page views and social followers. Measures attention, not value.
- Customer lifetime value. With a few months of history, any LTV figure is mostly a guess dressed as a calculation.
- Feature usage breakdowns. Useful when investigating a specific problem, noisy as a weekly number.
A one-page weekly scoreboard
| Metric | This week | Last week | Four weeks ago | Note | |---|---|---|---|---| | Accounts doing core job | | | | | | Activation rate (latest cohort) | | | | | | Month-2 retention (latest cohort) | | | | | | MRR and net new | | | | | | Runway (months) | | | | | | Customer conversations | | | | | | New customers via word of mouth | | | | |
Fill it in every Monday, in a spreadsheet, in under fifteen minutes. The note column matters most: write one sentence on why anything moved.
Getting early numbers from a real cohort
Before you have enough customers for percentages to mean much, the best source of signal is a small group of businesses using the product properly and telling you what happened. On LetsBeta, businesses apply to try your build, and each accepted Early Adopter files a mid-trial and end-of-trial report covering what worked, what broke, whether they would pay and a fair price. Those reports feed metrics 2, 3 and 6 directly. The success fee is 15 percent of what a converted customer pays in their first six months, then nothing, and you can see the plans on the pricing page.
To go deeper on the funnel view, read AARRR Pirate Metrics: Spot Growth Leaks in 90 Days. To connect these numbers to the bigger question they are all pointing at, see Product-Market Fit for Early SaaS: A Practical Playbook.
Seven numbers, one page, every Monday. Then go and have the conversations that explain them.