Ask a VP of Engineering what their org shipped last quarter and you get an answer in under a minute. Ask a VP of Design what their org contributed and you get a deck, assembled over a week, arguing from screenshots.
This is usually explained as a property of design: creative work is qualitative, impact is diffuse, you cannot put a number on taste. That explanation is comfortable and mostly wrong. Design is not intrinsically harder to measure. It is being measured against the wrong unit, by instruments built for a different kind of work.
The artifact is not the unit
The first instinct is to count artifacts. Files created, screens delivered, prototypes shipped, tickets closed. Every design organization that has tried this has abandoned it, and the reason is worth stating precisely: the artifact count is inversely correlated with the contribution that matters most.
The most valuable thing a senior designer does in a given month might be to kill a feature. The second most valuable might be to notice, in someone else's critique, that two teams are solving the same problem twice. Neither produces a file. Both produce more value than a week of screens.
Counting artifacts does not merely miss this work. It penalizes it. An organization that measures output and then asks its most experienced people to spend their time on direction, mentoring and systems has built a scoreboard that punishes the behavior it is asking for.
Attribution is the actual problem
The deeper issue is not measurement. It is attribution.
A shipped outcome has many parents. A conversion lift came from a redesign, which came from a research finding, which came from a support pattern somebody noticed, which was surfaced in a critique by a person who does not appear anywhere in the shipping record. By the time the outcome is measurable, the chain that produced it has been erased by the tools it passed through.
Engineering has an easier time here, and not because engineering work is simpler. It is because the chain is already written down. A commit references a ticket, the ticket references an epic, the epic references a goal, and the whole path survives in a system that recorded it as a side effect of doing the work. Nobody at any point had to remember to document the connection.
Design work generates a comparable trail. A Figma file, a Linear ticket, a research study, a Slack thread, a critique, a decision. What it does not generate is the links between them. Six tools each hold one piece, and none of them holds a relationship. The connective tissue exists only in the head of whoever was in the room.
That is why design impact feels unmeasurable. Not because the work is intangible, but because the evidence is real and unlinked.
What actually happens instead
In the absence of a trail, organizations fall back on the only mechanism available: memory, mediated by advocacy.
This is most visible in calibration. Practitioners of performance calibration describe the failure mode plainly: sessions run on memory and narrative rather than structured data, and the conversation defaults to advocacy, which rewards proximity and visibility rather than contribution. A designer whose manager is a strong verbal advocate does better than a designer of equal impact whose manager is not. Everyone in the room knows this is happening. Nobody has a better instrument.
The same substitution happens at the org level. A design leader asked to justify headcount reaches for whatever story is most available, because the number that would settle the argument does not exist. The zeroheight Design Systems Report 2026 found that 56 percent of design system teams name staffing as their single biggest challenge, and that only 23 percent agree they have adequate resources. Those teams are not failing to make the case because they are bad at arguing. They are making it without evidence, against functions that arrive with dashboards.
Three things that can be measured
Design impact is measurable if you change the unit from artifact to decision, and accept that a decision has a traceable path.
Where the capacity went. Not headcount, which is a budget line, but coverage: which squads had design, at what ratio, against which company priorities. Nielsen Norman Group puts the typical staffing at roughly one designer per twenty developers and one researcher per hundred, which tells you the median organization is thin enough that where those people are pointed is the single highest-leverage decision a design leader makes. Most cannot state it accurately for last quarter without rebuilding it by hand.
Where the work slowed and why. Cycle time from brief to shipped is measurable. So is the shape of the delay: the critique that never converged, the review that doubled, the handoff that sat, the rework a late requirement forced. This is the number that tells a design leader where to intervene, and it is the one most likely to already exist as timestamps nobody has assembled.
What the work changed. Not "design drove the lift," which nobody believes and which is usually unprovable. A traced path: this priority, these decisions, these people, this outcome. The claim is smaller than the one design organizations usually try to make, and unlike that one, it survives scrutiny.
Why this is worth fixing now
The urgency is new. AI made design output cheap, and the boundaries between design, research and engineering are dissolving. Figma's 2026 AI report found the share of developers doing design work rose from 44 to 60 percent in a single year, and 41 percent of practitioners say AI has meaningfully changed how their teams work together, against 7 percent two years earlier.
The same report found that output is measurably up while outcomes are not following. That gap is the entire problem in one sentence. A design organization producing more than it can account for is a design organization whose budget is decided by whoever tells the better story, and there is no reason to expect that story to favor design.
McKinsey's design index work found that top-quartile design performers achieved 32 percentage points higher revenue growth and 56 percentage points higher shareholder returns than industry peers over five years. That is the size of the prize for organizations that run design well. Almost none of them can currently demonstrate that they are one.
The shape of the fix
None of this is solved by adding a metrics dashboard on top of the existing tools. A dashboard is a view onto data somebody already structured, and the problem here is precisely that the data is unstructured and unlinked.
What it takes is a model of the organization: something that reads the work where it already happens, draws the links between a study and a file and a decision and a person, and keeps that record continuously rather than reconstructing it the week before a review. Once the links exist, the three measures above stop being research projects and start being queries.
Design is not unmeasurable. It has just been the last function anyone built an instrument for.