The Problem
Design has spent a decade measuring the wrong things.
Walk into any design org and the same metrics will show up on the wall: NPS, CSAT, usability scores, time on task. None of them are wrong, but none of them will save a team from a headcount cut. The CFO does not buy NPS. The CEO does not promote on usability scores. If you cannot translate design outcomes into the language the business already speaks, you will always be negotiating from a weaker seat.
The deeper problem is that most of these metrics are downstream of decisions design did not own. A usability score moves when engineering ships faster. NPS moves when pricing changes. Attaching design's credibility to numbers design cannot move on its own is how leaders end up holding the bag in a bad quarter for reasons that had nothing to do with the team's work.
Three Axes
Measure design on revenue, velocity, and trust. Nothing else.
Every design initiative my team runs has to ladder to one of those three axes. If it does not, we either reframe the initiative or stop doing it. That single discipline has done more for design credibility in my orgs than any framework, system, or research insight.
The reason these three work is that they map directly to what the rest of the executive team is already measured on. Revenue maps to the CFO and the CRO. Velocity maps to the CTO and the VP of Engineering. Trust maps to support, success, and brand. When design reports against the same three lenses, design stops being a translation problem in leadership reviews.
- Revenue: what shipped, what it moved, and what would not have moved without design owning the call.
- Velocity: cycle time from idea to shipped, how often the team learns, and how many experiments per quarter the team can run.
- Trust: customer signal (complaints, refunds, support load) plus internal signal (cross-functional satisfaction with design partnership).
What To Track
A small, defensible scoreboard beats a beautiful, ignored one.
I keep design's scoreboard to roughly six numbers. Two revenue lines we are accountable for. Two velocity numbers (cycle time and experiments shipped). Two trust numbers (top complaint volume and cross-functional NPS). Each number has a quarterly target. Each target ties to a named designer. The scoreboard lives in the same doc the PM team reads, not in a separate design dashboard.
The point is not to be exhaustive. The point is to be trusted. A small set of numbers, owned and reviewed every two weeks, beats a sprawling dashboard nobody opens.
A scoreboard that works has three properties. It is short enough that every designer on the team can recite it from memory. It is owned, with a name next to every number. And it is updated on the same cadence as the broader business, not on a separate design rhythm. The moment design's scoreboard is on a different calendar than the rest of the company, it stops being part of the conversation.
Design becomes uncuttable the moment it has its own revenue line.
How To Wire It Up
Instrument the work, not the team.
The biggest mistake leaders make when they try to introduce KPIs is to start with team metrics: utilization, output, throughput per designer. That is the path to a brittle culture and a brittle number. Instrument the work instead. For every shipped initiative, capture the hypothesis, the metric it was meant to move, the result, and the design decision that drove it. Six months in, you have a backlog of evidence you can point to in any forum.
Tooling matters less than the discipline. A shared notion or doc, updated by the designer who shipped, beats any analytics platform. The signal is in the writing, not the dashboard. Pair the qualitative log with one quantitative source of truth that the PM team already trusts, and you will never lose a debate about whether design moved the number.
The AI Layer
Add AI velocity KPIs before someone else does it for you.
In 2026, leaders are also being asked to show how AI is changing their team's output. Get ahead of it. Track concrete velocity gains: cycle time before and after AI tooling, percent of artifacts produced with AI assistance, hours reclaimed per designer per week. Share the numbers in review meetings before your CFO asks. Design leaders who report this proactively are the ones getting more headcount, not less.
Be honest about the numbers. If your team is genuinely 2x faster on certain workflows, say 2x. If a tool turned out to be a wash, say wash. The credibility you build by reporting the real numbers compounds. The credibility you lose by inflating them disappears the first time a curious exec actually looks under the hood.
Reviewing With Leadership
Bring three slides to every leadership review.
Three slides force discipline. They keep the review out of the weeds and they make it easy for the rest of the leadership team to engage. Most design reviews fail because they show too much work and too little judgment. The leaders who get more resources are the ones who walk in with a clear scoreboard, a clear bet, and a clear ask, every single time.
- Slide one: the scoreboard. Six numbers, last quarter vs this quarter, with a clear narrative on what moved and why.
- Slide two: the bets. What design owns this quarter, what we are betting will move, and what we will stop doing to make room.
- Slide three: the asks. The decisions, headcount, or air cover the team needs from leadership in the next 30 days.