How to set KPIs and goals for early-stage startups

Executive overview

Wrong KPIs send a startup in circles. The right primary metric gives you an objective read on health, a feedback loop on strategy, and a forcing function for prioritisation.

Pick one primary metric — revenue or active users — and pair it with three to five secondary metrics for a full picture. Set a weekly growth rate, track it honestly, and obsess over whatever blocks you from hitting it each week.

The primary metric is not a vanity signal — it must capture real, recurring value already delivered to the user.

Choosing the right primary metric

• Revenue is almost always the best choice — money paid means value already delivered.
• MRR (monthly recurring revenue) is ideal for SaaS: recurring payment proves ongoing value.
• Active users (DAU/WAU) are acceptable when audience scale is a prerequisite to monetisation (e.g. ad-supported businesses) or when strong network effects require a critical mass first.
• Avoid email signups or other leading indicators — they don't confirm value was actually delivered.
• Avoid MAU: a monthly cadence is too slow for early feedback and implies low usage frequency.
• Free users give different (weaker) feedback than paying users — charge early.

Four characteristics of a good primary metric

1. Quantifies value delivered — users signal value through money or time (revenue, DAU, WAU).
2. Captures recurring value — metric should reflect that users keep coming back.
3. Lagging indicator — value has already occurred, not just promised (payment beats signup).
4. Fast feedback loop — short enough cycle to inform decisions week over week.

Defining "active user" correctly

• Vague user counts are misleading — define exactly what counts as an active user.
• A user who returns daily for 10 minutes is fundamentally different from one who signed up once.
• For two-sided marketplaces, pick a metric that captures both sides getting value: Airbnb uses nights booked, Uber uses weekly trips.

Exception: biotech and hard tech

• If no regulatory barrier: use revenue in the form of paid contracts, LOIs, or POCs.
• If regulatory barriers prevent sales: primary metric becomes binary technical milestones — the minimum evidence needed to de-risk whether the science or tech works.

Secondary metrics

• Select three to five secondary metrics alongside the primary to get a 360-degree view.
• Example gap: tracking user growth without retention is dangerously incomplete.
• Optimising too many metrics at once causes analysis paralysis — keep the list short.

Setting growth goals

• The goal is a weekly growth rate — weekly cadence gives frequent feedback and breaks big targets into achievable steps.
• Benchmarks from recent YC batches: 5–10% week-over-week is good; 10%+ is exceptional; 20–50% month-over-month was common among strong performers.
• Small differences in weekly rate compound dramatically over a year — 1% WoW signals you haven't found product–market fit yet.
• Set your own goal based on what is ambitious and achievable for your specific product and market, not industry benchmarks alone.

Goal-setting guidelines

• Fast initial growth is normal when solving a real problem with latent demand — expect the rate to slow as volume grows.
• Account for time to sale: consumer is near-instant; enterprise can take months. Goals should reflect this early on.
• Prioritise organic growth over paid in the early stage — paid growth masks whether the product genuinely pulls users.
• Set exponential goals, not linear ones.

Two approaches to setting targets

• Growth rate method: pick a weekly rate, raise it if you consistently hit it, investigate urgently if you miss it.
• Time-boxed absolute goal: set a meaningful end-state (e.g. revenue or users at 10 weeks), back out the required weekly rate, and execute week by week.

Tracking and using your metric

• Print a forward-looking growth chart and update it weekly — this is what the Airbnb founders did.
• Each week, stack-rank your growth ideas and bet on the highest-impact one.
• Missing a week or two is fine as long as you know why.
• Key question every week: what is the single biggest obstacle to hitting my weekly target?
• If you don't know the answer, talk to more users — don't theorise in isolation.
• Consistently missing targets may mean you're working on the wrong problem.