— ESSAY · FOUNDER NOTE · 01
The unspoken cost of user research
Every founder is told: talk to your users. It is correct advice. It is also expensive in a way that the people who give it do not always remember. The cost is not the money. The cost is the lag between question and answer, and the way that lag distorts what gets shipped.
By the time a recruited panel has been scheduled, run, watched, and made sense of — three things have happened. The market moved. Your team built a workaround. And the question you started with is not the question that matters anymore.
— What the quarter without studies cost us
QUARTERS RUN THE OLD WAY
0
Recruited panels, paid incentives, awkward Zoom silences.
DECISIONS SHIPPED ON TIME
0%
From the one quarter we stopped running studies entirely.
SPEND COMPRESSED
0%
On the decisions that did not need a face — most of them.
— All figures from internal cadence. Nothing audited, nothing fabricated.
We tried, then we stopped
For the first two quarters of building Prior.Run we ran user research the right way. Recruited from the panel. Paid the incentives. Sat through the awkward Zoom silences. Took the notes.
What we learned was useful. What we learned was also six weeks late. And the most important things we learned were things we already suspected and could have tested faster a different way.
So we tried something heretical: stop running it for a quarter and see what broke.
Nothing broke. What broke was the comfort of saying we talked to users.
“The most expensive research is the kind that arrives after the decision is made.”
What we built instead
We built a synthetic audience we could ask any time, in minutes, with diverse profiles, and that produced reactions specific enough to act on. We calibrated it against the moments we did have real outcome data on, and we used the gap between the two as our signal for where the audience was still wrong.
The point was never to replace the highest-stakes conversations. The point was to compress the time between we are about to ship X and here is what the audience would feel. Weeks down to minutes is not an efficiency gain. It is a different way of designing.
What real research is still for
Pricing the unknown. Pricing the emotional. Pricing the thing where the only honest answer is to watch a face. We still do those. We just stopped pretending that all of our questions needed that ceremony.
The bar moved. The default moved. The expensive instrument is reserved for the questions that actually need it.
“Most of your design questions do not need a Zoom call. They need an audience that will tell you something true within the hour.”
— PRACTICE
What we tell teams now
Run synthetic first. Use the result to sharpen the question. If the synthetic answer is obvious, ship it. If the synthetic answer is contested across the audience, that is the question worth taking to real people.
User research is a precious resource. Stop spending it on questions whose answers were predictable.