The Metric That Matters Most - North Star vs Secondary vs Guardrail Metrics

Before we go deeper: what are these three metrics?

Before we get into how to identify each one, let's establish the basic idea.

When a product team measures performance, not every metric has the same job.

Some metrics tell us whether we're creating the value we set out to create.

Some tell us what is driving that outcome.

And some tell us whether we're improving one thing at the expense of something else.

Three distinct roles

Where are they used?

• Product strategy — deciding what the product should ultimately optimize for.
• OKRs & goal-setting — connecting company or product outcomes to measurable drivers.
• Product analytics — understanding which metrics deserve attention and which are merely diagnostic.
• Experimentation — defining the primary outcome of an experiment and the metrics that shouldn't deteriorate.
• Product reviews — creating a common language for discussing whether performance is improving and why.
• Executive dashboards — reducing dozens of metrics into a hierarchy that tells a coherent story.

The important thing is that these aren't three special types of metrics. They are three different jobs that metrics can perform.

A metric becomes a North Star, Secondary Metric, or Guardrail because of the role it plays in a particular product, team, or decision.

And once you understand that distinction, choosing the right metrics becomes much easier.

The moment dashboards become too successful

There is a moment in almost every product team when the dashboard becomes too successful.

More charts appear. More KPIs get added. Someone asks for one more metric. Then another.

Soon, the team has 47 metrics, 12 dashboards, and a weekly meeting where everyone looks at the same numbers but walks away with a different idea of what matters.

The problem is rarely a lack of data. It is a lack of hierarchy.

Good measurement is less about finding more metrics and more about knowing which metric plays which role.

That is where the idea of a North Star Metric, Secondary Metrics, and Guardrail Metrics becomes useful.

Think of them as three different jobs:

The North Star tells you where you're going. Secondary metrics tell you how you're getting there. Guardrails tell you what you must not break along the way.

Once you see metrics this way, a dashboard starts to look less like a collection of numbers and more like a map.

Product metrics work in much the same way.

A North Star Metric is essentially your destination. Secondary metrics are the signals that help you understand whether you're moving in the right direction. Guardrails make sure you don't drive off a cliff while trying to get there.

The mistake many teams make is treating all three as if they are the same thing. They aren't.

1. The North Star Metric: Where are we trying to create value?

Let's begin with the most misunderstood one.

A North Star Metric is not simply the biggest number on your dashboard. It is not necessarily revenue. It is not necessarily users. It is not necessarily orders. And it certainly shouldn't be chosen because someone senior happens to like it.

The North Star should represent the value your product consistently creates for its users, in a way that also connects to the long-term health of the business. That's an important distinction.

That is why choosing a North Star requires something more fundamental than asking: 'What number do we want to increase?' Ask instead: 'What value do we want to create repeatedly?' That's a much better starting point.

A useful test for your North Star

1. Does it represent user value? If the number goes up, does that generally mean users are getting more value?
2. Is it connected to business value? A metric can be great for users and terrible for the business, or vice versa. The North Star should sit somewhere in the overlap.
3. Can the product team influence it? A metric that nobody can meaningfully influence is a poor North Star.
4. Does it encourage the right behavior? This is perhaps the most important question. Because what you measure eventually influences what people do.
5. Can it survive beyond one quarter? A good North Star shouldn't change every time the quarterly OKRs change. It should represent something deeper.

2. Secondary Metrics: How do we get there?

Once you know where you're going, the next question becomes: 'What drives the North Star?' This is where secondary metrics enter.

Suppose your North Star is Successful Orders. Now ask what has to happen before a successful order occurs. Perhaps: App Open → Menu View → Add to Cart → Checkout → Payment Success → Order.

Suddenly, your secondary metrics become much more meaningful.

• Menu-to-cart conversion
• Cart-to-checkout conversion
• Checkout-to-order conversion
• Payment success rate
• Average time to order
• Repeat order rate

These metrics explain the machinery underneath your North Star. And this is where product analytics becomes particularly powerful.

The North Star tells you that something happened. Secondary metrics help you understand why.

Imagine successful orders decline by 4%. That number tells you there is a problem. But it doesn't tell you where the problem lives. You investigate further. Menu views are stable. Add-to-cart rate is stable. Checkout starts are stable. Payment success rate has dropped sharply. Now you have a direction.

The North Star is the outcome. Secondary metrics are the levers and leading indicators.

A secondary metric is useful because it helps explain or influence the North Star. It becomes dangerous when the team forgets why it mattered in the first place. The metric becomes the goal. And once the metric becomes the goal, people get very creative.

3. Guardrail Metrics: What must we not break?

Now we arrive at the third category. And personally, I think guardrails are one of the most underappreciated ideas in product measurement.

Imagine a product manager says: 'Let's make checkout faster.' Sounds good. The team runs an experiment. Checkout completion improves by 6%. Everyone celebrates. Then someone checks fraud. It has increased by 15%. Another person checks customer complaints. They've increased too.

The experiment succeeded according to one metric. It failed according to the product.

This is why you need guardrails.

A guardrail answers a different question: 'What are we unwilling to sacrifice to improve the primary outcome?'

Example hierarchy: Food delivery

The goal isn't simply to increase orders. It is to increase orders without destroying the experience or economics that make those orders valuable.

A good product team needs all three. Without a destination, you wander. Without a dashboard, you can't navigate effectively. Without warning lights, you can reach the destination with a broken car.

The framework: Outcome → Drivers → Constraints

North Star = Outcome | Secondary Metrics = Drivers | Guardrails = Constraints

Example: Music streaming product

Now the metric hierarchy becomes obvious. You're not saying: 'Increase everything.' You're saying: 'Increase the outcome, understand the drivers, and protect the constraints.' That is a much healthier way to build a product.

How to actually identify them

For Airbnb, it might be something close to: 'Our product creates value when guests successfully find and stay in a place they want.' For a food-delivery product: 'Our product creates value when customers successfully receive food they want.' For a learning product: 'Our product creates value when learners make meaningful progress.'

The exact wording will differ. That's okay. The exercise forces you to think about value before measurement. Only then should you start choosing metrics.

Step 1: Find the value moment

Ask: What is the moment when the user actually receives the value? Not when they open the app. Not when they click a button. Not when they see an ad. The moment when the product has genuinely done its job.

Call this the Value Moment. Your North Star should be closely connected to this moment.

Step 2: Find the drivers

Now work backwards. Ask: 'What needs to happen for the Value Moment to occur?' Break the journey into stages.

For example: Discovery → Consideration → Action → Value.

At each stage, identify the metrics that influence movement to the next. These become your potential secondary metrics.

Notice the word potential. Not every useful metric deserves a permanent place on the executive dashboard. Sometimes a metric is useful for diagnosis but doesn't need to become a KPI. That distinction alone can save a team from dashboard inflation.

Step 3: Find the things you refuse to break

Now ask the uncomfortable question: 'If we aggressively optimize the North Star, what could get worse?' This is where you search for guardrails.

• Orders increase but cancellations increase.
• Engagement increases but retention falls.
• Conversion increases but refunds increase.
• Revenue increases but customer satisfaction falls.
• Time spent increases but users accomplish less.

Every product has trade-offs. Guardrails make those trade-offs visible.

Step 4: Test the hierarchy

Then ask one simple question: 'If this metric improves, what does it tell us?' If the answer is unclear, you may not know what role the metric is playing. And that's usually a sign that the metric needs another look.

The experiment test

There is another powerful way to use this framework. Imagine you're running an A/B test. Your experiment changes the checkout experience. Your primary metric: Order Conversion +4%. Looks great.

But then: Refund Rate +8%. That's a problem.

Now imagine: Order Conversion +4%, Refund Rate flat, Cancellation Rate flat, Customer complaints flat. Now you have a much more compelling result.

The experiment didn't simply move a number. It improved the outcome without violating the constraints.

Did we actually make the product better, or did we simply find a way to make one metric look better?

One metric can play different roles

There's an important nuance here. Metrics don't have permanent identities. The same metric can be a North Star for one team and a secondary metric for another.

For example: Revenue could be the North Star for a business team. But for a product team, it might be a downstream outcome.

Similarly: Daily Active Users might be a North Star for one product, but merely a diagnostic metric for another.

The role of the metric depends on the decision you're trying to make and the value you're trying to create.

So don't ask: 'Is revenue a North Star metric?' Ask: 'Is revenue the right outcome metric for this product and this team?' That's a much better question.

The 3 questions I use

• What are we trying to improve? → North Star
• What can we change to improve it? → Secondary Metrics
• What must remain healthy while we do it? → Guardrails

That's it. You don't need a 40-page KPI taxonomy to start. You need clarity about outcomes, drivers, and constraints.

A final thought

There is something beautiful about a good metric framework.

It reduces complexity without pretending that complexity doesn't exist.

A product is complicated. Customers are complicated. Businesses are complicated. But our measurement system doesn't need to mirror all of that complexity. It needs to help us make better decisions.

The best dashboards I've seen don't have hundreds of numbers. They have a small number of numbers that answer important questions.

Where are we going?

What's moving us there?

What are we risking along the way?

That's the real purpose of a metric hierarchy.

Because ultimately, metrics are not the destination. They are the language we use to decide whether we're building something worth building.

And perhaps that's the most important metric of all.