You make these calls in the dark.
You ship a product you believe in, then you start guessing. You read about a refund on Monday and cannot tell whether the formula missed or the person quit too early. You watch a loyal customer churn and never learn they were two weeks from feeling it.
The work you got into this for slowly turns into decisions you make blind. The alternative used to be a study you would run once and shelve.
The analysis plan is locked and date-stamped before anyone enrolls.
The complete distribution is reported, every customer measured against their own starting point.
Our fee does not move with your result. We are paid the same either way.
Your customers already know. Now you can say so.
We measure the people who actually bought your product, each against their own starting point, and hand you the exact sentence their answers support. Not a number and a shrug. The wording, the disclosure that rides next to it, and the file that defends it.
Every part of it, explained →The figure. Fixed in the analysis plan before anyone enrolled.
The sample. Who the sentence is about, and who it is not.
The verb. Perception, not effect. The word that keeps you on the right side of the line.
The outcome. In your customer's language, from the cleared vocabulary.
The timepoint. The week response shows up, not the week your guarantee ends.
The anchor. The phrase nobody on your shelf can write, because they never met the customer before she started.
The disclosure, already written and sized for the surface.
Self-reported results reflecting within-person change over time, not a controlled clinical trial. Measurement commissioned by [your brand]. Individual experiences vary.
A cofactor you can publish raises the number.
Add a condition your customer controls and your claim gets stronger and more defensible at once, because it names who the number is about. Add one we only estimated, and it stays inside.
The dashboard is yours to see, all of it, the misses included. What you may publish is the cleared set. We draw the line before you do.
See the exact sentences your own customers would support, on one of your products, before any call.
See your demoWhy this is different.
The usual study tests a panel of strangers once and hands you a marketing line. We measure your own customers, continuously, and show you who responds, when, and why.
| A claims study | Rationally Well | |
|---|---|---|
| Who is measured | A panel of matched strangers | Your own customers |
| Who is in the denominator | The people still taking it | Everyone who enrolled at purchase, the quitters included |
| How often | Once, then it ages | Continuously, refreshed as buyers arrive |
| What you get | One average, one claim | Who, when, and why, plus the sentence you can publish |
| The deliverable | A line for the label, then shelved | A living read you act on |
| The incentive | Paid to produce a sellable claim | Published pricing that does not move with your result |
People respond differently, and that is documented.
People respond differently to the same nutrient for reasons that are documented in the published literature. We read each customer against their own baseline and report the spread.
Response can depend on where you start.
In a two-year randomized trial, a B-vitamin regimen was associated with a slower rate of brain atrophy, concentrated in people with high baseline homocysteine (about 53% lower rate in that subgroup).
Response can depend on who you are.
In a randomized omega-3 trial, how much reached the brain differed by genotype: the rise in brain EPA was roughly three times greater in people without the APOE4 variant than in carriers.
These are cognition studies. We include them as documented examples that response can vary by baseline and by genotype. They are not evidence about sleep, supplements, or any product on this page. The point is narrow: people differ, and the difference is measurable.
What brands ask first.
What exactly do we walk away with? +
How do you make money? Do you take a cut of our sales? +
Is this IRB reviewed? +
What happens if the data comes back unflattering? +
Do you promise results, or a seal? +
Does the companion coach my customers? Does that change the read? +
Put a number on retention.
You set the improvement you want to model, and we do the math on your own numbers. Prefilled with Example figures, change any field to use yours.
You set this. It is your hypothesis to model, not our promise.
Published, and it does not move with your result.
No call required to see it.
Beyond 300 active testers the retainer scales per active tester, set with you as you grow. The full plans page →