prior.runFrom the lab.
Research notes, methodology deep-dives, and the thinking behind our engine. Written by the people doing the work, not the marketing team.
Recent dispatches.
Featured
We tested our AI personas against Nobel-Prize psychology. One was off by a single percentage point.
A behavioral validation report. Seven classic experiments. Every result published — including the misses.
live test
Cursor's Pricing Page Loses 80% of SaaS Buyers in Under Five Minutes
Cursor.com/pricing converts the developers it targets — and silently bleeds every adjacent buyer who lands on it.
mood drop
Robinhood: The Trust Ledger Never Stopped Bleeding
Retail brokerage discourse keeps circling back to one January, one halt, and a permanent suspicion.
live test
Substack pricing lost 100% of B2B buyers in under five minutes
Substack's pricing surface failed every SaaS buyer we sent — hidden fees and mismatched social proof killed the click.
comparison
Prior.Run vs. UserTesting: When to Pick Which
Depth instrument vs. frequency instrument. The compare is the answer.
live test
Robinhood Lost 80% of Skeptics Before the Scroll.
Premium aesthetic. No social proof. Four compliance flags. The synthetic audience verdict.
synthetic users
Why Synthetic Users Give Better Design Feedback Than Real Ones
The counterintuitive case for simulated audience reactions.
LLM bias
How We Built a Skeptical Audience (On Purpose)
LLMs are polite by default. That's a disaster for design feedback.
synthetic users
Our Audience Has Read Every Pricing Page You Will Ever Build
What happens when synthetic users develop pattern recognition.
synthetic users
Why Context Changes Everything in Design Feedback
Two identical people. The same pricing page. Opposite reactions. Both right.
SaaS
SaaS Pricing Page Patterns: What Actually Works
What happens when skeptical, trusting, and anxious users see the same pricing page.
experimentation
Kill Your Losing Experiments Before You Build Them
Most A/B tests fail. What if you could predict which ones before writing a line of code?
design review
Design Reviews Are Theater
And what to do instead of another meeting.
design validation
User Research Without Users
A practical guide to making better design decisions faster.
design comparison
How to Compare Designs When Your Team Can't Agree
A framework for moving from preferences to decisions.
AI design review
Every AI Design Tool Agrees With You
And how to use them without getting burned.
live test
Affirm Lost 87% of Fintech Users to Pricing Opacity
Affirm's brand carries weight, but hidden cost structure and seven competing CTAs fracture intent at the hero.
live test
Apple One lost 86% of visitors to hidden pricing
Apple's brand pulls users in. Missing prices and zero social proof push most back out.
live test
Amazon Prime lost 83% of high-intent shoppers in under three minutes
Amazon Prime's landing page has the bones — but zero social proof and a four-plan maze stall the majority.
live test
Google Workspace Lost 89% of B2B Buyers in Under Three Minutes
Google Workspace converts on brand alone — but missing social proof costs the page 89% of skeptical B2B buyers.
live test
Meta Quest lost 87% of shoppers to a single missing element
meta.com/quest establishes the product fast, but shoppers stall waiting for reviews, ratings, and real evidence.
live test
SoFi Lost 84% of Fintech Users to CTA Ambiguity
SoFi's legitimacy signals land, but unclear button promises stall the majority of high-intent visitors.
live test
Airbnb's homepage lost 86% of shoppers in 136 seconds
Airbnb's brand opens the door, but missing prices and reviews kept shoppers from walking through it.
live test
Netflix's hero converts the bold — but 83% scroll past unsure
Netflix's homepage anchors price but hides the next step, costing it most of the room.
live test
Coinbase lost 95% of fintech users in 129 seconds
Coinbase.com has the brand, the prices, and the offer — but no proof, so the audience walks.
live test
DoorDash converted 13%, lost 87% to missing trust signals
DoorDash leads with a $0 delivery hero, but pricing opacity and zero social proof bleed conversions.