Kill 'The Team Would Be Slower Without It' as the Test of a Leader's Real Work

Patrick Neeman's piece on UX Collective this morning wants design leaders making again, and most of it I would hand to a design director without a note. One sentence in it goes on the Kill List, because it describes the leader a small team cannot afford.

The short version

Neeman's argument is the player-coach: Bill Russell coached the Celtics three seasons while playing center and won two championships, design leadership grew on cheap money and the correction is a hangover, and the metric now rewards output per person, so the leader has to bring their own hands. His numbers, as he reports them: the share of SMB workers who manage anyone fell from just under 25% in 2019 to about 15% by late 2024 per Gusto, Google cut 35% of managers over teams under three, and Figma's 2026 report has designers in development rising from 21% to 41% in a year. His test for whether the leader's piece was real work: if the team would be slower without it, it was. Kill that test. It selects for the leader on the critical path, the person Marco Di Santo described last week as the place every correction waits. The replacement test is that the next ten outputs shipped against the piece with the leader out of the room, and the piece that passes is a constraint set: the worst output you would still ship, what the product refuses, and a rubric with 20 real cases. The number to report is the percentage of live surfaces with a current constraint set.

What he gets right, which is most of it

The case is built on the money. Peter Merholz's design inflation, as Neeman reports him: about fifteen years of near-zero rates, Capital One from roughly 90 designers in 2014 to 500 or 600 within a few years, and a leadership model where team size was the business case. The pandemic pushed hiring further and Meta let go of more than 11,000 people. Much of what is blamed on AI, he says, is that hangover, and he has Oxford Economics attributing about 55,000 job losses to AI in the first eleven months of 2025, under 5% of reported layoffs. The cuts are the correction; AI shapes what the smaller teams look like afterward.

Then the shape. Andy Jassy's September 2024 memo asked Amazon's senior leaders to raise the ratio of individual contributors to managers by at least 15%, with the reason "most decisions we make are two-way doors." Gusto's number on SMB managers. Google's 35%. The super individual contributor, Catt Small's staff designer with team leadership plus in-the-pixels work, which Neeman says he is and is proud of. Figma's 2025 report with developers at 59% and designers at 31% on AI for core work, and Lütke's memo making reflexive AI use a baseline at Shopify with "everyone means everyone."

The section I would sign outright is the one on craft. Figma's State of the Designer 2026, as he reports it, had 58% of 906 designers defining craft as visual polish, and polish is the part models now produce fastest. Then: an AI feature has thousands of possible outputs, no designer will see most of them, someone has to decide what a good answer looks like, write it down as examples and evaluation criteria, and keep checking real outputs against it. Teams call those evaluations, and they are becoming the specification. His example is a contract summary feature: ten real agreements, the five facts a lawyer needs from each, and a rule that the summary never invents a renewal date. That is The Rubric Is the Spec in one paragraph, and it is the best thing in the piece.

The sentence that goes on the list

Then the action item. Ship something monthly: own one real piece of production work each month, a component or a prompt, and put it through the same review as everyone else's. And the test for whether it counted: "if the team would be slower without the leader's piece, it was real work."

Kill that test.

It selects for the leader on the critical path. Kill the Correction Queue argued with Marco Di Santo's post from last week on exactly this person: someone who built the agent fast because they held requirements, build, and release, and then became the place every correction waited. That person passes Neeman's test every single week. The team is slower without them. That is the problem, not the proof.

Look at the two pieces he names. A component the team would be slower without is a dependency with a title on it; the day the leader is in a board meeting, the feature waits. A prompt is the customer layer, in the terms the FDE series uses: changed the same day, no review, versioned, and stale with the next model. A leader whose monthly piece is a prompt has made themselves the maintainer of the one artifact that changes most often.

Both pieces are real making. Neither is leadership making, because neither scales past the leader's calendar, and the calendar is the constraint his whole piece is about.

The replacement test

The piece was real if the next ten outputs shipped against it with the leader out of the room.

You Cannot Mock a Distribution has the case. At Smartcat everyone could agree on a perfect translation in five minutes. We spent a month fighting over the worst one we would auto-send without a human. That fight was the product. The month was the leader making, in Neeman's sense, and what it produced was not a component or a prompt. It was a line: the worst acceptable output, written down well enough that two people would sort ten real outputs the same way. The auto-send then ran against that line on every job nobody reviewed, with no one waiting on me.

Design's Kill List has the same move on a ritual rather than a surface. At Salesforce I killed the weekly review everyone quietly hated and loudly defended: same slot, new format, both running for two weeks, and the condition that would bring the old one back said out loud. Nothing ever did. That was a leader's hands on a ritual, and the test it passed was that the team ran the new format without me in it.

So the monthly piece that passes is a constraint set for one live surface. Three columns from the distribution chapter: the best case worth aiming at, the worst case you would still ship, and what the product refuses to do. Plus a rubric, in the sizes Figma's Trial Closed 70% of the PM Gap. The Gate Is the Rubric, Not the Designer. already uses: three to five criteria, at least 20 real cases, graded pass or fail per dimension, calibrated with two people on five outputs before anyone calls it real. Neeman's contract summary paragraph is this artifact. He has the right piece and the wrong test for it.

The number to report

Neeman's four action items end with "price one feature," "ship something monthly," "sort the doors," and "rewrite one ladder." Good list. The number underneath it is the one the design handbook's Team Shape(coming Jan 20) chapter says almost nobody can state: what percentage of your live surfaces have a current constraint set.

That is the player-coach's scoreboard. Not what the leader shipped, which is the headcount metric with the sign flipped, but how much of the product runs on judgment that was written down and therefore keeps operating on surfaces nobody is looking at that week. Written judgment is the multiplier. A leader who adds one surface a month to that percentage is making, in public, through the same review as everyone else, and the team gets faster without them every time.

Under Product Leadership the claim is that leadership work now is deciding which coordination roles stop being necessary and saying so out loud. Neeman is saying it out loud, and the title that stops being necessary is the one whose only evidence is a calendar. The one that stays is the leader who can state the percentage.

This week: pick the live surface your team touches most and write its worst-acceptable column, with ten real outputs sorted against it by two people. If they sort the same way, that is your monthly piece. Then compute the percentage for the whole product, and put it in the next leadership update where the headcount number used to be.

Related answer: What should a design leader build with their own hands each month?

Sources: Patrick Neeman, "Smaller teams want design leaders who make, again. Here's how.," UX Collective, October 28, 2026, read in full; Peter Merholz, Andy Jassy, Tobi Lütke, Catt Small, Gusto, Oxford Economics, and the three Figma reports are cited only as Neeman reports them, and none was opened for this post. Marco Di Santo, "The Person Who Built Your AI Agent Is Now Its Bottleneck," Operational AI, October 20, 2026, as argued with in "Kill the Correction Queue" (October 20, 2026). Falk Gottlob, "Team Shape(coming Jan 20)," "The Rubric Is the Spec," "You Cannot Mock a Distribution," and "Design's Kill List" in The Design Operating Model, and "Figma's Trial Closed 70% of the PM Gap. The Gate Is the Rubric, Not the Designer." (October 19, 2026).