KPIs That Matter

Most founders are tracking the wrong things. Not because they're lazy — because the metrics that are easiest to measure are almost never the ones that matter most.

You've got a dashboard full of numbers. Monthly active users. Social followers. App downloads. Page views. You open it every morning, feel a little buzz when the lines go up, and call it "data-driven." But if none of those numbers are forcing a decision, you're not doing analytics. You're doing therapy.

The hard truth: vanity metrics are comfortable, and real KPIs are uncomfortable. Let's fix your dashboard.

The Vanity Metric Trap

Here's how you know a metric is vanity: removing it from your dashboard would change nothing about how you run your business.

Downloads don't tell you if anyone's coming back. Page views don't tell you if anyone's buying. Follower counts don't tell you if anyone cares. These numbers feel like progress because they move — but they don't move your business.

Pinterest famously ignored raw sign-ups for years and tracked something far more specific: whether a new user had pinned at least one thing within their first session. That single behavioral signal was far more predictive of long-term retention than total registered accounts. One number drove product decisions. The other just looked good in press releases.

Ask yourself: if this metric doubles tomorrow, does it change what I do on Monday? If the answer is no, cut it.

Leading vs. Lagging: The Distinction That Changes Everything

Most founders track lagging indicators — revenue, churn, profit. These tell you what already happened. By the time a lagging metric looks bad, you've already lost weeks you can't get back.

Leading indicators predict future outcomes. They're what you can actually act on.

Lagging: Monthly revenue
Leading: Number of qualified demos booked this week
Lagging: Annual churn rate
Leading: % of users who complete onboarding in 7 days
Lagging: Net Promoter Score
Leading: Time-to-first-value in the product

Stripe didn't just track total payment volume. Early on, the team obsessed over how long it took a new developer to make their first successful API call. That leading indicator — time-to-activation — predicted everything downstream: retention, expansion, referrals. They optimized that number relentlessly, and the lagging metrics followed.

Track what predicts the outcome, not just the outcome itself.

The question isn't "how did last month go?" It's "what's telling me how next month will go?"

How Investors Read Your Metrics (vs. How You Do)

When you show a deck to an a16z partner or a YC demo day judge, they're not impressed by revenue in isolation. They're reverse-engineering your business model from your metrics.

Here's what they're actually looking for:

CAC vs. LTV ratio — Is this business fundamentally unit-economic? They want to see LTV at least 3x CAC, ideally higher.
Payback period — How long until you've recovered what you spent to acquire a customer? Over 18 months is a red flag in most SaaS.
Net Revenue Retention (NRR) — Are existing customers expanding? 120%+ NRR means your revenue compounds even if you stop acquiring new customers.
Activation rate — What % of sign-ups actually experience the core value of your product?
Growth rate consistency — Not just the number, but whether the trajectory is compounding or decelerating.

You might be proud that you hit $50K MRR. An investor is thinking: what's the churn? What's the expansion rate? What does NRR look like? Is this $50K compounding or leaking?

The metric you're celebrating might be the one they're most skeptical of. Come prepared with the full picture.

The Right 3–5 Metrics for Your Stage

There's no universal set of KPIs. The metrics that matter for a pre-product startup are completely different from a Series B company scaling a sales team. Stage determines signal.

Pre-product / 0 to 1:

Weekly active user growth in beta (are early users returning?)
Qualitative NPS from your first 10 users (would they be very disappointed if this went away?)
Time to first core action (how fast are you getting people to the "aha moment"?)

Early traction / Post-launch:

Activation rate (% who reach value within X days)
Week-1 and Week-4 retention cohorts
CAC by channel (what's actually working, and at what cost?)

Growth stage / Scaling:

NRR / Net Dollar Retention
Sales cycle length (especially for B2B)
CAC:LTV ratio by segment
Revenue per employee (operational efficiency signal)

Uber in its early days wasn't tracking total rides. They tracked rides per driver per day — a single efficiency metric that told them whether supply and demand were balanced in each city. One number. Massive leverage.

The goal isn't more metrics. The goal is fewer, better ones.

Building a Decision-Forcing Dashboard

A good KPI dashboard does one thing: it tells you what to do next. If you open it and nothing changes about your priorities, it's decoration.

Here's how to build one that actually works:

Start with your one north star metric — the single number most correlated with long-term business success. For Airbnb it was nights booked. For Slack it was messages sent within a team. What's yours?
Pick 2–3 input metrics that you believe drive the north star — these are what your team actually influences week to week.
Add 1–2 health metrics to catch problems early — churn rate, support ticket volume, error rate. These shouldn't be goals; they're guardrails.
Kill everything else. Seriously. If it's not in those five slots, it shouldn't be on the primary dashboard.

Google's early team famously tracked queries per day as their north star. Everything else — page rank, index size, crawler efficiency — was an input metric. One number unified the entire company's priorities.

The companies that scale cleanly are almost always the ones where every employee can tell you the north star metric and whether they're on track this week.

Common KPI Mistakes Founders Make

Even founders who understand this framework still get it wrong. Here's where things go sideways:

Changing metrics too often. You can't build intuition for a number you've only tracked for 6 weeks. Commit.
Averaging instead of cohort-ing. Average retention hides massive variance between customer segments. Always break it down by cohort, acquisition channel, or use case.
Tracking what's easy, not what matters. If a critical metric is hard to instrument, that's a signal — not an excuse.
Using metrics as performance theater. KPIs shown to investors should be the same ones your team looks at every Monday. If they're different, something's wrong.
Ignoring counter-metrics. Every primary metric needs a counter-metric. Optimizing activation rate at the expense of user quality is a trap. Track both.

Final Thought

The founders who build durable companies aren't the ones with the best dashboards. They're the ones who figured out — usually the hard way — which single number tells them the truth about their business, and then built every process, every hire, every product decision around moving it.

Pick fewer metrics. Make sure they hurt a little to look at when things are off. If your KPIs are always green, you're either crushing it or you're measuring the wrong things.

Most of the time, it's the second one.