Every article about activation is written for a company that already has a data team. It assumes you have thousands of users, an event taxonomy someone maintains, and enough traffic for a cohort chart to mean something. Then it tells you a healthy SaaS activation rate is somewhere around 25–30%, as though that number is useful to someone with forty signups.

It isn’t. Not yet. But activation still matters at forty signups — arguably it matters more, because it’s the only thing that tells you whether the product works before the revenue does.

This is the version of that article for people who launched last week.

I build LaunchPulse, a $5/month analytics tool, so I have an obvious interest in you caring about this. I’ll flag it whenever it’s relevant. Most of what follows is about the concept, and works the same whether you measure it with my product, someone else’s, or a SQL query against your own database.

The three numbers, in order

There are exactly three things happening on your site, and people conflate them constantly.

Acquisition. Someone showed up. You measure it with visitors and traffic sources.

Conversion. Someone signed up. Visitors who became users.

Activation. Someone got the thing they came for. A user who reached the moment your product is actually for.

The gap between conversion and activation is where most early products quietly die. You can have a great signup rate and a dead product, and traffic dashboards will show you nothing but green. People arrive, register, poke at an empty state, and never come back. Acquisition looks fine. Conversion looks fine. Nobody ever got any value, and you find out three months later when nobody renews.

Activation is the first number that measures whether your product does something for people, rather than whether your landing page is persuasive.

What “activation” actually means

Activation is a user reaching a milestone that signals they’ve experienced your product’s core value. That’s the standard definition and it’s fine, but it’s abstract enough to be useless without a concrete choice behind it.

The concrete choice is: which single event, when a user does it, means they got what they came for?

Some real examples:

Product	Signup means	Activation means
Project tool	account created	first project created and shared
Invoicing app	account created	first invoice actually sent
Waitlist	joined the list	invite accepted, first login
Analytics tool	account created	first real event received from their site
Store	account created	first order placed

Notice that in every row, activation is something the user does themselves, and it’s the thing that would make them sad if you took it away.

The famous examples, and why you should be careful with them

You’ll run into these within ten minutes of researching activation: Facebook’s “7 friends in 10 days.” Slack’s “2,000 team messages.” They get repeated as though they were derived from careful mathematics.

They mostly weren’t. Andrew Chen — who was close enough to that world to know — has pointed out that Facebook’s number could just as easily have been “10 friends in 12 days” or “five friends in one day.” Mixpanel’s team put it more bluntly: magic numbers are “a very useful illusion.” The number gets rounded to something memorable so a company can rally around it, not because 7 is statistically distinct from 6 or 8.

There’s a second, more serious problem with the folklore: correlation is not causation. Users who added 7 friends were more likely to stick around. That does not mean pushing an indifferent user to add 7 friends makes them stick around. It may just mean people who already liked Facebook added friends. If you chase the metric without the underlying value, you build growth hacks that move a number and change nothing real.

So use these stories as a shape — “find the action that means someone got value, and drive toward it” — and don’t go looking for your own magic constant. You don’t have the data to find one, and it probably wouldn’t be causal if you did.

Why a percentage is the wrong metric at your scale

Here’s the part nobody tells early founders.

Activation rate is a percentage, and percentages need volume to mean anything. For a sense of scale: Convert, an A/B testing company, advises running a test only if you can send around 10,000 visitors to each variant, and recording roughly 1,000 conversions before treating the result as conclusive. You have forty signups. Your activation rate is going to swing from 20% to 40% because two people did or didn’t click something on a Tuesday.

This is also why A/B testing is a waste of your time right now, and why “our conversion rate dropped 15% this week” is usually noise wearing a suit.

It gets worse if you’re on Google Analytics. GA4 applies data thresholding: when the number of users in a report is small, it withholds the data entirely to protect privacy. Google won’t publish the exact cutoff — the official docs only say thresholds are “system defined” and can’t be adjusted — but practitioners consistently report it kicking in somewhere around 40–50 users. So at precisely the stage when every single signup matters, the most popular analytics tool in the world may show you an empty card and a note about minimum aggregation thresholds.

The tool isn’t broken. It was built for a different problem than yours.

What to use instead: time to value

When you can’t trust a rate, measure a duration.

Time to value is how long it takes a user to get from signup to their first-value moment. Median, not average — one person who activated after three weeks shouldn’t drag the whole picture.

It works at small numbers because it doesn’t need a denominator. With five activated users you have five real durations. That’s not a statistically significant sample, but it’s not noise either: it’s five true facts about how long your onboarding takes. If four of them took two minutes and one took nine days, you have something specific to go look at, and you can go ask that person what happened. That conversation is worth more than any dashboard at this stage.

Three things time-to-value tells you that a rate won’t:

Whether your onboarding has a wall in it. A median of six days for a product that should activate in five minutes means something is blocking people.
Whether the activation event is even the right one. If everyone hits it within ten seconds of signing up, it’s too easy to be meaningful — you’re measuring a UI step, not value.
Where to spend your next week. Long time-to-value is an onboarding problem. Short time-to-value with bad retention is a product problem. Those need completely different fixes.

Picking your first-value event

Four tests. A good first-value event passes all of them.

They do it themselves. Not something you do to them. “Received welcome email” is not activation.

They’d notice if it were gone. If removing the feature wouldn’t bother them, reaching it didn’t mean anything.

It’s not automatic. If every signup gets it by default, it’s a step, not a milestone. A trial that starts automatically at signup is a good example: it will have a 100% rate forever and tell you nothing.

It’s not too far down the funnel. If it takes three weeks and a sales call, you won’t get enough of them to learn anything this quarter.

The most common mistake is picking something too close to signup because it’s easy to instrument. “Completed onboarding” usually means “clicked through four modals,” which correlates with nothing. Pick the moment the product actually did its job — even if it’s harder to track.

Instrumenting it, concretely

Whatever tool you use, activation needs three things:

1. A signup event. Fired when the account is genuinely created — after the backend confirms, not when the form submits. Form-submit events count failures as conversions.

2. A first-value event. Fired when the user reaches the milestone you picked. Prefer firing this from your backend after the operation completes, for the same reason.

3. A stable user identifier on both. This is the one people miss, and it silently breaks everything.

That third point deserves its own paragraph, because it’s the single most common reason an activation number stays at zero. Analytics tools group events by actor. If you never tell the tool who the user is, each browser is an anonymous visitor. Someone signs up on their laptop and reaches first value a week later on their phone, and those look like two unrelated strangers. Activation can’t be computed at all — not inaccurately, but at all.

In most SDKs the call looks like this:

// as soon as a session exists, on every page load — not only at signup
lp.identify(user.id); // your internal id, never an email

Call it on every page load where someone is signed in, not once at signup. And use your internal database id, never an email address or anything else personal.

We shipped this bug in our own product, incidentally. We tracked signups and first value correctly, never called identify(), and stared at a dashboard reading “0 reached first value” while the events were plainly visible in the log. The events were fine. The join was impossible.

What to ignore for now

Things that are genuinely not worth your attention at launch scale:

Bounce rate. Ill-defined and unactionable when a “bounce” might be three people.
Session duration. Long sessions can mean engagement or confusion. You can’t tell which from a number.
A/B tests. You don’t have the volume. You’ll be reading noise and shipping the winner of a coin flip.
Any dashboard with more than about six numbers on it. More metrics at this stage means more opportunities to fool yourself.
Week-over-week percentage changes on tiny numbers. Going from 3 signups to 5 is not “up 67%.”

Track 3–5 things. Look at them weekly, not hourly.

What to do this week

Write down, in one sentence, what a user must do to have gotten value from your product. If you can’t, that’s the real finding — and it’s worth more than any analytics setup.
Instrument that single event, fired from the backend, with a stable user id.
Look at the durations, not the rate. Five real numbers beat a meaningless percentage.
Find the person with the longest time-to-value and ask them what got in the way.

Step four is the one people skip, and at forty users it’s the highest-value thing on the list. Your analytics tell you that something is wrong and roughly where. A five-minute conversation tells you why. At this stage you have few enough users that talking to them is still possible — that’s an advantage you lose permanently, so use it while you have it.

FAQ

What’s a good activation rate? Published benchmarks tend to put a healthy SaaS activation rate somewhere around 25–30%, but treat that as trivia until you have thousands of users — and note that these figures are usually quoted without a study behind them. Below a few hundred, your own trend over time is the only comparison worth making.

Should conversion be signup or payment? Signup, in most cases. Keep payment as a separate event you also track. If you make payment your conversion, everything above it in the funnel becomes invisible, and the top of your funnel is where the fixable problems live.

My trial starts automatically at signup. Should that be a funnel stage? No. If it happens for 100% of signups automatically, it carries no information — it’ll be a permanent 100% step. Track the moments where users actually make a choice.

Can I do this without an analytics tool? Yes. Two timestamps in your users table — signed_up_at and first_value_at — and a SQL query gets you median time-to-value. That is a completely legitimate starting point, and better than an unconfigured dashboard.

How long until this data is useful? Time-to-value is useful with your first five activated users. Rates need hundreds. Cohort retention curves need thousands and months.

Related reading:

Conversions and activation — configuring both stages
Identifying users — visitors, users, and why the id matters
Tracking events — sending custom events
The Pulse dashboard — exact metric definitions

About the author. I’m Ignacio Van Droogenbroeck. I build LaunchPulse, analytics for products that just shipped, at $5/month. I built it because every analytics tool I tried was designed for a company with a growth team, and I kept wanting something that answered three questions — is anyone coming, is anyone signing up, is anyone getting value — without a taxonomy meeting. If that’s your situation too, the docs start here.

Sources: the 25–30% activation benchmark as commonly quoted by PayPro Global (note: widely repeated, rarely sourced to a study); the magic-number critique from Mixpanel and Andrew Chen; GA4 thresholding from Google’s own documentation; A/B test sample-size guidance from Convert; the case for tracking few metrics early from PostHog.