Build the Strategy Stack Before You Argue About Metrics

Most product teams argue about metrics before they know what they are building toward. The OKRs feel random. The roadmap feels like a list someone made under pressure. The vision sits in a slide deck nobody opens. The fix is not a better dashboard. It is a stack: vision at the top, strategy in the middle, metrics and tactics below, with each layer feeding the one above it. When leaders skip the stack and start with numbers, they end up managing things that do not matter. Here is how to build it in the right order.

The deep cut

• The metric flows from the vision, not the dashboard. Sebastian Straube warns that Daily Active Users is a vanity count with no tie to customer value.
• A number with a deadline is an OKR, not a North Star. The Strategy-Metrics Sandwich shows teams who blur the two end up with random quarterly goals.
• Reverse-engineer leading indicators from your best users. Netflix tracked new members who added three titles to their queue in session one.

Start at the top and work down

A product strategy has layers, and they only work in order. Both Ravi Mehta and Roman Pichler lay the same point out visually: vision sits on top, strategy in the middle, roadmap and tactics at the bottom. The job is to check that every layer exists and that they connect.

The common failure is a gap you cannot see from inside one layer. A team with a sharp roadmap and no vision is just busy. A team with a grand vision and no OKRs cannot ship. Before you touch any single quarter's plan, walk the whole stack and name the weak layer. That is the one to fix first.

Vision answers three questions, not one

Vision is not a mood. Gibson Biddle's DHM model, written up by Bandan Singh, turns it into three plain questions: how will the product delight customers, what makes it hard to copy, and what business model experiments keep it profitable. Each year, pick five or six strategic bets and test whether they actually hit D, H, or M.

Biddle also pushes leaders to think in phases over ten to fifteen years. His GLEe model reads Netflix's vision as three steps: get big on DVDs, lead downloading, expand worldwide. You do not do all three at once. You sequence them. That phasing is what keeps a long vision from becoming a wish list, and it tells your team what to ignore right now.

Pick a North Star that measures customer value

Once the vision is set, the metric should almost fall out of it. Straube's rule is strict: a North Star metric measures the moment a customer gets real value, not the moment your ego does. Views on a banner and registered users are vanity counts. They tell you nothing about what people value.

For Netflix, the vision is "entertain the world," so the North Star becomes time spent watching, built from number of users times hours per session times sessions per week. That breakdown matters. A good North Star is one number made of a few driving ones, so you can run an A/B test on a recommendation algorithm and see it move. If your current top metric is revenue or sign-ups, you are measuring what the company wants, not what the customer gets.

The sandwich: strategy connects the number to the quarter

North Star metrics and OKRs are different animals, and mixing them causes the mess. The Product Strategy-Metrics Sandwich puts it cleanly: North Star metrics are persistent goals that shape strategy, while OKRs are point-in-time goals that carry it out. A quick test: if there is a date attached to the figure, it is an OKR. If it takes more than one team to move, it is a North Star.

Strategy is the filling in the middle. It turns the big number into clear choices, and those choices become proxy metrics that answer "how would we know we won." The catch is that proxy metrics lag. They move slowly and no single team controls them, which makes them useless for week-over-week work.

Turn lagging proxies into leading indicators

For teams to act, you need leading indicators, the signals that move fast and respond to what one team does. The sandwich method says to reverse-engineer them from your prioritized customers: find the patterns your successful users share, then track those. That is exactly what the Netflix queue metric does. Retention is the lagging truth; "three titles added in the first session" is the leading signal a team can chase this week.

Matt Hicks offers a primer on leading versus lagging indicators worth handing to any PM before they mix metrics on a roadmap. The judgment call is trading certainty for speed. Lagging metrics are more certain but tell you too late. Leading ones are fuzzier but let you steer. You want both: lagging proxies for annual OKRs, leading indicators for the daily work.

Three questions for your team

• Walk our stack top to bottom: where is the gap, a vision with no metric, a strategy with no OKRs, or OKRs with no strategy behind them?
• Does our North Star measure the moment a customer gets value, or is it a number that just makes us feel good?
• For this quarter's goals, which are leading indicators one team can move, and which are lagging proxies we should only report up annually?