The Factorization of UXR: The K-shaped Divide Underneath Modern Research Work

Late-night thoughts. The kind you write because the picture has been forming in the background for a while and tonight it finally sat still long enough to look at.

UXR is not one field anymore. It has not been one field for a while, but the labels held long enough that most people did not notice. One side is collapsing in real time. The other is busier and more leveraged than it has ever been. They share a job title and almost nothing else.

I call this the factorization of UXR. The word does double duty. Factory floor: the work has been broken into stations that can be optimized, automated, or outsourced one at a time. Factors of production: researchers themselves are treated as inputs to a process. Interchangeable. Measured by throughput. Swappable in principle and increasingly in practice.

The factorization of UX research already happened. We're living in the result of it. The shape is mostly set.

The factory came for research

If you think about it UXR has been broken into stations. Intake, recruit, moderate, transcribe, synthesize, deck, close. Each station SLA'd. Each one optimized on its own. Each one with a vendor ready to take it over the second the math works.

This is Taylorism applied to research. It's been creeping for a decade. AI made it cheap enough at scale that finance noticed, and once finance notices something it stops being a question of whether and becomes a question of how fast.

Inside that factory the researcher is a node. The team is a throughput function. The product is a deliverable. Whether the deliverable changed any decision is, for the purposes of the factory, an externality.

The K is already there

Not every company runs the factory the same way. The companies sorted themselves into two arms by now, and where a company sits on the K determines how research actually lives inside it.

On the upper arm of the K, research informs decisions. Findings change calls. Senior researchers are in the conversation while bets are being framed. The output has consequence.

On the lower arm, research gets produced, consumed, and archived. Decisions happen on other inputs. Gut. Last quarter's number. Whoever was loudest in the staff meeting. The research is documentary. It tells the story of what was already going to happen.

Both arms have researchers. Both run programs. Both ship decks. What separates them is downstream, and it shows up late.

Nobody designed this

Most companies did not pick their arm. They accumulated one.

I've written about organizational accumulation before. The short version is that UXR functions inside most companies were not designed at all. They were assembled. One ad-hoc hire at a time. Scope set by whoever asked first. Team built around whoever happened to be around. Frameworks imported from wherever the latest senior hire came from. Researchers brought their last org's playbook because nobody told them to do otherwise.

The arm a company sits on now is the consequence of that accretion. It was not a strategic choice. It was not a manifesto. It was a long series of small decisions, and none of those decisions was actually about what kind of research function the company should have.

The exceptions are the obvious ones. Google, Meta, a handful of companies that designed the function deliberately when they built it. They sit closer to the upper arm because they built parts of the function more deliberately than most. Even those companies have lower-arm pockets. The brand does not iron the arm out across the org. Everyone else is wherever the accumulation deposited them. The accumulation, statistically, deposited most of them lower than they think.

Nobody knows what arm they are on

This is what makes the whole thing painful, and it is why a lot of you reading this are nodding without being entirely sure why.

The theater of research is identical in both arms.

Same intake forms. Same kickoffs. Same fieldwork. Same readouts. Same decks. Same stakeholders saying the right things about being user-centered and evidence-based and letting the data guide us. Same research roadmaps in Notion. Same ResOps stack. Same vocabulary, fluently performed, in every meeting.

From the inside, the day-to-day on the lower arm and the day-to-day on the upper arm look essentially the same. The signal that distinguishes them, which is whether any decision actually moved because of the work, lives downstream of you, arrives late, and is buried in noise. By the time you can read it cleanly you have already spent two years there.

I've sat in readouts where the deck got presented, the findings landed, the stakeholders nodded, and watched the actual call get reversed in a Slack DM later that afternoon. Nobody in the room thought of that as the research not landing. The readout happened. The deck existed. The theater was complete.

Researchers do not know what arm they are on. The companies do not know either. They have researchers. They have a program. They reference findings in slides. They believe they are an upper-arm company because they have the apparatus of one.

Almost no company actually runs the test that would answer the question. The test is straightforward. Count the decisions that changed because of research over the last twelve months. Divide by the studies shipped in that period. Look at the ratio. Most companies would rather not.

The K is inside companies too

The K does not only sort companies. It sorts teams inside companies, sometimes within the same building, sometimes within the same VP's org.

A researcher on the wrong product surface at a famous brand can be doing factory work while a researcher two floors away is shaping bets. The deck looks the same. The badge looks the same. The work does not. Meta, Google, the rest of the usual suspects all have lower-arm pockets the brand absorbs without anyone naming them.

The inverse happens too. A small team at a mid-tier company can be doing genuinely upper-arm research because the team has made a project, over years, of bending its corner of the org toward something better. This is rare and slow. It does not show up in any survey of the state of UXR. It is visible to the few people doing it and the few people they are pushing.

The honest question is not only which company you work for. It is which team, under which VP, on which product surface, with which set of stakeholders. The same badge can mean very different work depending on where in the org you actually land.

What the upper arm actually looks like

On the upper arm, decisions reference research findings by name, in writing. Senior leaders ask research questions before scoping work, not after they've already built the thing. Research and product roadmaps are coordinated rather than maintained as parallel artifacts by people who occasionally email each other. Researchers are in the room when bets are being framed. The team is small relative to the company, senior-heavy, and paid like the strategy function it is. AI shows up as leverage inside the function, compressing cycle time and expanding the surface area one senior person can cover, not as a substitute for the judgment work.

That is what it looks like when it is working. I have seen pieces of this and the whole thing in a few places. The whole thing is rarer than people pretend. Way rarer.

What the lower arm actually looks like

On the lower arm, research is requested after the decision is effectively made (though everyone often pretends otherwise). The role is risk reduction or stakeholder cover. Findings get received, thanked, and not acted on, because nobody has built the mechanism that would convert insight into action. The team is large relative to its influence. Junior-mid-heavy. Qualitative leaning. High throughput. Low to some autonomy. Work is scoped from a ticket (or a roadmapping exercise, or a dm) rather than from a question. AI gets used to increase throughput. The factory gets faster. It does not get smarter. Velocity goes up. Judgment does not. Stakeholders speak fluent research. The theater is polished. The decisions are made elsewhere, and this is not a secret to anyone who works there.

Another way to see it is to look at the ceiling. In a lower-arm org, no researcher has ever moved above a certain level. The Director slot might exist on paper. The Principal track might be written into the ladder. Nobody from research actually crosses into the space where research is shaping the direction of the company. The ceiling is invisible and absolute. If you cannot name a senior researcher at your company who has the ear of the CEO, you are not on the upper arm, regardless of what brand sits on your business card.

The trap built into that ceiling is that titles on the lower arm come easier than they do on the upper arm. The factory needs Seniors and Staffs and Principals to manage throughput, and the factory has rungs that need bodies standing on them. A Principal Researcher at a lower-arm company can have the same title as an upper-arm Senior with none of the access an upper-arm Senior actually has. This is how people get stuck. The title was the part the company could hand you. The shape of the work was not.

No shade in any of that. The K is not a moral ranking. It is a description of what is there.

AI did not hit the K evenly

AI did not hit UXR evenly. It hit the lower arm of the K much harder than it hit the upper one. The asymmetry is structural, and once you see it you cannot unsee it.

Researchers on the lower arm have felt this harder than anyone else. Headcount pressure landed there first. Layoffs landed there first. Scope erosion landed there first. If you've been wondering why your particular corner of UXR feels like it's collapsing while a friend at a different company sounds fine, that's where to look. You and your friend are probably not on the same arm of the K.

The upper arm felt AI too. The impact is real. It is different in shape. AI compresses cycle time on the upper arm. It expands the surface area one senior researcher can cover. It does not replace judgment work, because judgment work was never the bottleneck. Velocity was. Senior researchers on the upper arm are more leveraged now than they were two years ago.

I think about this asymmetry a lot. AI did not create the K. It widened it. The middle of the K, the place a company could plausibly hold a few years ago, is mostly gone. The companies that used to sit in the middle have been pulled toward one arm or the other, and most of them have been pulled lower.

The transition trap

Moving from the lower arm to the upper arm is hard. It is harder now than it was three years ago, and I think it's going to keep getting harder.

Lower-arm work does not build the muscles that upper-arm hiring screens for. The factory rewards throughput, project clearance, deliverable speed. The upper arm hires for framing ambiguous questions, owning a strategic narrative, operating without a script, changing a senior leader's mind. Five years on a factory floor and your portfolio is a list of studies you completed. Five years on the upper arm and your portfolio is a list of decisions you changed.

The hiring loop reads the difference even when the JD does not say it out loud. You can feel it in the kinds of follow-up questions you get. Tell me about a time you changed someone's mind. Tell me about a piece of work that did not get used and what you did about it. Tell me about a study you killed. The factory does not generate good answers to questions like those, because the factory is set up to keep running, not to make those kinds of bets.

People do cross. I know researchers who have done it. The ones I know who crossed did it on the side. Case studies built from work they had to wedge outside their actual job, getting senior at a lower arm company first and using that as the bridge into an upper arm one. None of them rode the internal escalator their employer was offering. The escalator was going the other way. If it existed at all.

Most of the career pain I see in this field right now lives in exactly this gap. I do not think it has been fully named yet. People can feel the friction without being able to articulate the shape of it. That is part of why I'm writing this on a Tuesday night instead of going to sleep.

What this changes for researchers

The career question stops being where you want to work and becomes which arm you are on, and how you would know.

You answer it with downstream evidence. Studies shipped tell you almost nothing. Decisions changed tell you almost everything. Whether you were in the room when bets got framed tells you the rest. JDs are usually honest in a way nobody intended them to be. The verbs do all the work. Execute, deliver, support, contribute to, partner with, those all live on one arm. Shape, frame, drive, define, set the direction for, those live on the other. The pay bands are different. They have always been different.

Your company is not going to do this exercise for you. Your company does not know which arm it is on, so it cannot tell you. This is the kind of exercise you do for yourself, on a night like the one I'm writing this on, with no audience.

If you are on the upper arm, AI is a multiplier and the competitive pressure is from other senior researchers who got fluent with it earlier than you did. The risk is not the model. The risk is the person at the next company who has been using it for two years already.

If you are on the lower arm and you want to stay, the play is to become the person who runs the factory rather than the person being run by it. The factory needs operators. The operators get paid differently than the nodes.

If you are on the lower arm and you want to cross, the path I have seen actually work is mostly off-hours. Side projects. Public writing. Case studies built around the kinds of artifact the upper arm hires for. The current employer is not going to develop you into something it does not value. It cannot. The surface to do it on is not there.

Where this leaves me

The K is legible now. For a long time the lower arm got told to itself as a younger version of the upper arm, the immature version that would get there once it matured. That story has run out. I do not think there is a maturation path from one arm to the other. They are two different functions, paying differently, doing different jobs, with different futures.

What I do not know is what this means for any of us five years out. The upper arm gets smaller and more leveraged. The lower arm gets larger, cheaper, and more replaceable around its edges. The middle is where almost everyone would prefer to live and where almost nobody is going to be allowed to stay. I have written this whole piece and I am not entirely sure where my own situation sits on a five-year horizon, which is probably the most honest thing I can put on my blog right now.

🎯 Subscribe for more of this. The frequency depends on how often something annoys me enough to write about it. Lately it has been often.