Why “North Star Metrics” Aren’t Enough
Every product team loves a clean dashboard.
One big number at the top. Green arrows. A sense that we’re moving in the right direction.
That number is usually called the North Star Metric.
And it’s useful.
But it’s also dangerous.
Because most teams mistake visibility for understanding.
The North Star gives direction — not control
A North Star Metric tells you where you want to go, not why you’re moving or what’s breaking along the way.
If your North Star is Monthly Active Users, Revenue, Retention, or Engagement — you’re already late to the problem.
Those numbers are outcomes, not behaviors. They move after something has already gone right — or wrong.
When MAU drops, the damage has already happened. When revenue spikes, you often don’t know which lever caused it.
Teams celebrate or panic... without learning.
The Slack lesson nobody talks about
In 2015, Slack was growing explosively. Their North Star — Daily Active Users — looked incredible. But internally, the team noticed something troubling: new teams were signing up in waves, yet many went silent within two weeks.
The DAU number kept climbing because new signups masked the quiet churn underneath.
Stewart Butterfield’s team didn’t just watch the North Star. They dug into a specific input metric: messages sent per team in the first 72 hours. They found that teams sending fewer than 2,000 messages in their first month had a 93% chance of churning.
That single insight reshaped their entire onboarding. They stopped optimizing for signups and started engineering for early momentum — prompts to invite teammates, guided channel creation, integrations that generated automatic activity.
The North Star kept rising. But now they understood why.
The real work happens below the North Star
High-performing teams don’t ask: “Is the North Star going up?”
They ask: “What behaviors are pulling it up — and which ones are quietly killing it?”
That means tracking supporting metrics: activation steps, time-to-value, feature adoption by cohort, error rates, drop-off moments, and trust signals like refunds, complaints, and manual overrides.
These are messy. They don’t fit on a single slide. They don’t look good in exec decks.
But they tell the truth.
Why North Stars fail in the real world
North Star obsession creates three common failure modes:
Local optimization. Teams juice the metric instead of the product. Facebook learned this painfully when engagement-maximizing algorithms drove time-on-site up while user wellbeing and trust eroded. The metric was green. The product was rotting.
Blind spots. In 2019, a mid-sized SaaS company I advised watched MRR climb 40% year-over-year. Leadership was thrilled. But buried in the data: NPS had dropped from 45 to 12, support tickets had tripled, and their largest cohort was quietly shifting to annual plans to “lock in pricing before switching.” Revenue was a lagging indicator of a customer base preparing to leave. By the time MRR reflected the damage, rebuilding trust took 18 months.
Decision paralysis. When the number doesn’t move, no one knows what to change — because nothing upstream is visible. I’ve watched teams stare at flat retention curves for quarters, running random experiments because they couldn’t identify which specific moment in the user journey was leaking value.
The product becomes reactive instead of intentional.
A better way to think about metrics
Keep the North Star. But treat it like a compass, not a steering wheel.
Your real operating system should be:
1 North Star (direction)
3–5 input metrics (behavior) — and these should be specific enough that a single team can own each one
Weekly learning loops, not just reporting — the question isn’t “what happened” but “what did we learn and what are we trying next”
Here’s the test: pick your North Star and ask your team to explain, in one sentence, which user action most directly moves it. If you get five different answers, you don’t have a metric problem. You have a clarity problem.
And clarity is what great products are built on.