Somewhere in a founder's Notion, there's a metrics dashboard that gets looked at exactly once a week — during the Monday standup — and then never informs a single decision. I've built that dashboard. More than once. The KPIs were right there, neatly coloured, and completely useless, because a metric without a target is just a number with good posture.
Setting KPI targets for startup growth is where most teams fall apart. Picking the metrics is the easy bit; almost every guide will hand you the same list of CAC, LTV, churn and runway. Then you have to say out loud: "Our target is X by month Y," and suddenly everyone gets nervous, because a target is a promise you can miss. That's the part worth writing about.
Key Takeaways
- A KPI target is a decision, not a wish — anchor it to a business constraint (runway, capacity, a hiring plan) rather than a round number you like.
- Work backwards: fix the outcome and the deadline, then divide into monthly steps and check the step against your current trajectory.
- Two or three targets per team beat twelve. The 80/20 rule here is about attention, not about which metrics exist.
- Write down the assumption behind every target. When the assumption breaks, the target should change — that's not failure, that's the system working.
- Give targets a floor and a ceiling. A single number invites gaming; a range invites honesty.
How to set KPI targets that actually drive startup growth
The trap is setting targets in the same meeting where you pick the metrics. Those are two different jobs. Picking a KPI is an analytical exercise; setting a target is a forecasting exercise, and forecasting badly is how you get a team that quietly stops believing in the dashboard.
Start from the constraint, not the ambition
Here's the thing: a target has to come from somewhere real, or it's fiction. In my experience the honest anchors are almost always one of these:
- Runway. You have a known number of months of cash. That caps how long a growth target has to prove itself.
- Capacity. Your support team handles a certain volume of tickets per person per week. Double the customer count and you either hire or break something.
- A committed plan. A hiring plan, a fundraise timeline, a product launch date that sales is already promising to prospects.
Notice what's absent: "we should grow because growth is good." That's not an anchor. That's a vibe.
My rule, and I'll defend it: if you can't name the constraint that makes a target matter, don't set the target yet. You'll end up chasing a number that nobody can explain six months later, and when it's missed, the post-mortem turns into a debate about whether the number was ever right instead of what actually went wrong.
Work backwards from the outcome
Take a concrete case. Suppose the constraint is a plan to raise in roughly four quarters, and the thing investors will ask about is predictable recurring revenue. You know today's monthly recurring revenue. You know the rough multiple of monthly growth that makes a story credible at that stage (ask two or three founders one step ahead of you; you'll get a range, not a law).
Now do the arithmetic that everybody skips: divide the gap by the number of months, and compare that monthly increment to what you've actually added in each of the last three months. If the required increment is far above your recent average, you don't have a target. You have a hypothesis that something must change — a new channel, a price change, a second salesperson — and that hypothesis is the real work.
Spoiler alert: the first pass usually produces a number that makes everyone laugh nervously. Good. That reaction is information. Either the timeline moves, or the growth mechanism changes, or the number was never the point.
Turn the target into monthly steps you can check
A quarterly target is too coarse to manage. If you only look at it at the end of the quarter, you learn about a miss when it's already too late to react. Break it down so that each month has a checkpoint — and, critically, so that a bad month triggers a conversation in week two rather than week eleven.
- Set the outcome target and its deadline.
- Divide into the monthly increments required.
- For each month, name the leading indicator that would show up first if you were on track (a demo count, a trial activation rate, an inbound reply rate).
- Add a floor: the point below which you re-plan rather than push harder.
That floor is the piece almost everyone forgets. Without it, a bad month just gets absorbed into "we'll catch up next month," which is a sentence people say right up until the runway conversation.
How to set up KPI targets, step by step
If you want a repeatable procedure you can run with your team in an afternoon, this is roughly what I'd do. It's not glamorous. It works.
A four-step procedure
Step one: name the constraint. Write one sentence: "The reason we need this target is ___." If the sentence is hard to finish, stop and finish it.
Step two: pick two or three KPIs, no more. One revenue or growth metric, one efficiency metric (what it costs to acquire or serve), one health metric (retention, churn, satisfaction). Anything beyond that is reporting, not managing.
Step three: write the target as a range with a date. Not "improve retention" and not "churn of exactly 1.8%." Something like: "Monthly churn between X and Y by the end of Q3, with a floor of Z below which we re-plan." A range forces you to think about what's acceptable, not just what's aspirational.
Step four: attach the assumption. Every target rests on something you believe about the world — a channel will keep working, a conversion rate will hold, a hire will ramp in a certain number of weeks. Write that belief down next to the number.
Here's what that looks like in a messy real table, which is far more useful than a tidy one:
| KPI | Why it's in the target set | How the target is framed | Assumption to revisit |
|---|---|---|---|
| New recurring revenue per month | Directly tied to the fundraise constraint | A range, checked monthly | That the current channel keeps producing at the same rate |
| Cost to acquire a customer | Efficiency — protects the runway | A ceiling, not a goal | That ad prices and conversion rates stay roughly stable |
| Customer retention | Health — makes growth compound | A floor plus a stretch | That onboarding hasn't changed |
| Runway in months | The constraint itself | A minimum, reviewed whenever spend changes | That hiring happens on the planned schedule |
Note the asymmetry. Some targets are ceilings, some are floors, some are ranges. A single "target number" for everything is one of those habits that feels rigorous and isn't.
The review cadence matters more than the number
A target that gets reviewed quarterly is decoration. The teams I've seen do this well look at the numbers weekly, but only change the targets monthly or quarterly, which stops the constant goalpost-shifting that drives people nuts. Weekly look, monthly learn, quarterly re-plan. That rhythm is boring, and boring is the point.
What is the 80/20 rule for startups?
The 80/20 rule, in its original form, is the observation that a small share of inputs tends to produce a large share of outputs — a handful of customers generating most of the revenue, a couple of channels producing most of the signups, a few features driving most of the usage. In a startup context it's used as a discipline: spend most of your attention on the few things that move the outcome, and let the rest run at a lower standard.
Applied honestly, the rule isn't about which metrics exist. It's about which ones get a target at all. Most startups have a long list of things they could measure and a very short list of things they can actually influence this month. The 80/20 question to ask is simple: of everything on this dashboard, which two or three, if they moved, would change our situation?
Sometimes the answer is uncomfortable. A team I worked with was tracking a dozen marketing metrics while the actual constraint was that the salesperson had no capacity for more demos. The 80/20 answer was "hiring and ramping a second rep," which is not a metric at all. That's fine. If the highest-leverage lever isn't a number, the target should be the hire.
What are some important KPIs for a startup?
The short answer: the ones that map to how your business actually makes money. The longer answer depends on your model, but a few categories come up again and again.
The four categories worth knowing
- Growth: new customers, new recurring revenue, activation rate for new signups. The activation one is underrated — a trial that never activates is a leak you're paying to fill.
- Efficiency: cost to acquire a customer, and how long it takes to earn that cost back. The second number is the one that bites you during a lean stretch, because a channel can look great on acquisition cost and still drain cash for months.
- Retention: churn, repeat purchase rate, expansion within existing accounts. For subscription businesses, retention quietly determines whether growth compounds or just replaces what leaks out.
- Cash: burn and runway. Ugly numbers, but they're the constraint most targets should be anchored to.
On the ratio of lifetime value to acquisition cost: you'll see a lot of confident numbers quoted online, and I'd treat all of them with suspicion, including any I might quote. The honest version is that the "right" ratio depends on your margin, your sales cycle and how much cash you have to fund the gap. A ratio that's healthy for a high-margin software business is not the same as one for a company shipping physical goods. Work it out from your own cash cycle rather than from a benchmark someone posted.
Not every number deserves a target
Follower counts, page views, brand awareness, "engagement." These aren't useless — they can be leading indicators — but they're a poor basis for a target because they don't connect cleanly to a business outcome. Early on, I set a target for social reach, hit it, and learned precisely nothing about whether the company was healthier. Total waste of a quarter's attention.
Is it true that 90% of startups fail?
Not in the tidy way it's usually stated. The "nine out of ten" figure gets repeated everywhere, but the underlying data is messier than the headline: it depends heavily on how you define a startup, when you start counting, and what counts as failure — a shutdown, an acquisition below the money raised, or a founder simply walking away. Different data sources produce different survival curves, and the widely quoted figure is more of a folk statistic than a precise measurement.
What I'd take from it is directional rather than numerical: a large share of new businesses don't survive their early years, and running out of cash is one of the most common reasons. Which is exactly why KPI targets anchored to runway beat targets anchored to ambition. The point of a target isn't to hit it. It's to find out early — while you still have options — that you're not going to.
Two questions I get asked a lot about targets
"What if we miss?" Miss once and you adjust the plan. Miss twice in the same direction and the target was wrong, not the team. Say so out loud; the credibility you lose by pretending otherwise is worse.
"Can targets be private?" Some can, but the ones that drive growth usually can't. A target nobody outside the founding team knows about is a target nobody is helping you hit.
The dashboard I mentioned at the start — the one that got looked at once a week — I eventually deleted most of it. Three numbers stayed, each with a range, a date, and a sentence underneath explaining what had to be true for the target to make sense. That sentence did more for our decisions than any of the charts. If you take one thing from this, make it that: a target without its assumption written down is just a number with good posture, and it will let you down quietly.