Rationally Wellby Hinge Testlab
For supplement brands · founding access

A study gives you a claim. We show you who responded, and when.

Most research hands you one number to put on the label. Rationally Well reads your real customers and shows who responds, when, and what travels with the result, so you understand every hit and miss, not just the average.

Preview your responder picture No outcome promises, ever · founding partners lock their terms
What a claims study leaves you
★★★★
92% reported improvement Example
One number for the label. It cannot tell you what made it work, or what got in the way.
Same customers, read honestly
What Rationally Well shows
Glycine alongside
+12%
Low baseline vitamin D
+8.4%
Late caffeine
−9%
Low adherence
−14%
What travels with a stronger or weaker read · the conditions alongside each hit and miss Example
Right now

You are guessing, and it costs you.

You ship a product you believe in. Then you start guessing. You guess at who it is for in the ad account. 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. You change the formula and hold your breath. The work you got into this for slowly turns into decisions you make in the dark.

The ad budget you scaled to the wrong people, because nothing told you who responds.
The refund you read about Monday, weeks after the data would have told you quietly.
The five-star reviewer you could have built a case study around, if you had known who they were.
The loyal customer who quit at week three, two weeks from the effect they would have felt.
And the one without a number: the nights you spend guessing, because the alternative was a study you would shelve. You pay that one in sleep.
Try it yourself

Preview your responder picture in ten seconds.

Pick a category, pick what you want to understand, pick a cohort size. We will mock up the kind of readout you would walk away with. Not a claim to paste, a picture to act on.

1 · Category
2 · What you want to understand
3 · Cohort size
1,000 customers
Sample readout Example

46%31%23%
Four responder phenotypes. Strongest group, low magnesium, clean baseline, near 84%.
WKS 6 TO 8
Glycine alongside+12%
Late caffeine−9%
Confidence level at this size: higher
Sample readout, illustrative. Your study, your customers, your numbers.
On your own surface

Here is your product page, the day after one read.

Yesterday it was a star rating and a soft claim. Today you know who responded, when, and why. Same product, same customers, read honestly.

Today
Magnesium for sleep
★★★★★4.2 · 1,000 reviews
"May support restful sleep."
one blended score
no idea who it worked for
no idea what tipped the result
no read on your actual customers
With Rationally Well
Magnesium for sleep Example
46%31%23%
Four responder types, observed on your customers.
who responds, as phenotypes you can name
when each group starts to feel it
the cofactors that travel with a hit or a miss
read continuously, on your own buyers

We do not hand you a number to paste and a seal to slap on. We hand you a living read of your customers.

How the data gets there

Your customers get a companion that supports follow-through.

The reads on this page come from an app in your customer's hand. It does more than collect data. It is built to support follow-through through the window where people usually quit, and shows them their own progress. Follow-through is where retention is won or lost, and the companion is built to support it.

9:41
Your own pattern
Start → now · last 47 nights
Time to fall asleepIllustrative
You are falling asleep sooner than when you started
26 min
your recent nights, across this window
Your start
41 min
Now
26 min

About 15 min sooner than your own starting point.

Night wakings
1.7 → 0.6
per night · start → now
Morning energy
2.4 → 3.5
of 5 · start → now
How your pattern is logged
47
nights logged
38
nights took Stillwell

Logging this consistently is what makes your pattern readable. It is a record of what you did, not part of the change shown above.

It captures the data.

Baseline at the start, then short check-ins. The source of every read on this page.

It keeps them going.

Reminders and encouragement through the weeks 6 to 8 window, when people quit just before the effect they would have felt shows up.

It shows them their own trend.

Their response reflected against their own baseline, where they started and where they are now. Their own data, not advice on what to take.

And we say so. Because the companion actively supports follow-through, your read reflects the real-world experience, your product plus the support around it, which is what your customer actually lives. We label it that way every time. The honesty is what keeps the read worth trusting.
Why this is different

A claim you shelve, or understanding that compounds.

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 what travels with it.

A claims studyRationally Well
Who is measuredA panel of matched strangersYour own customers
Your customersNever contactedSupported through a companion
How oftenOnce, then it agesContinuously, refreshed
What you getOne average, one claimWho, when, and why
The deliverableA line for the label, then shelvedA living read you act on
The incentiveDesigned to produce a sellable claimDesigned to report the result, flattering or not
What you walk away with

A living read, not a one-time certificate.

Concrete deliverables, refreshed as your customers keep reporting.

A live responder dashboard

Cohort response, retention through the first window, and the trend, on your own customers.

A phenotype map

Who responds, as named groups with their traits, so you know who to feature and acquire.

A cofactor read

What travels with a result: what is taken alongside, baseline status, adherence. The pattern alongside each hit or miss.

Insight refreshed over time

Each finding observational, with its confidence level and limits stated plainly.

On purpose, no seal. We do not sell a badge or a guaranteed claim. A seal is a one-time stamp. Honest, ongoing understanding is the asset, and reporting the read exactly as it comes back is what makes it worth trusting.
The science

Why a single average is the wrong unit.

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.

Documented · responder by baseline

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).

Smith et al., PLoS One, 2010 (VITACOG). DOI
Documented · responder by genotype

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.

Arellanes et al. (Yassine lab), EBioMedicine, 2020. DOI

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.

Promises

What we won't do.

We won't promise outcomes.

We help you measure and report real results. We will not promise a share will respond or that the data will flatter you.

We won't dress up the result.

If a read comes back unflattering, you get it with its limits intact. The honesty is the point.

We won't sell you a claim.

Our fee is for honest measurement. We report what we find, whether or not it flatters the product.

We won't lock you in.

Your data, your customers, your brand, yours. Read continuously for as long as it serves you.

Who this is for

The brands this fits.

If a brand you respect lives here, that is the conversation we want.

Subscription and DTC supplement brands

where retention in the first weeks is the whole economic model.

Products where customers struggle to tell if it is working

and an honest responder picture would settle the question.

Brands that outgrew a one-time claim study

and want understanding that keeps reading, not a line they already shelved.

Brands living through a formula change

that want to know what moved, and for whom, before the market tells them.

If two or three names came to mind, an introduction is the most useful thing you could do.

Common questions

What brands ask first.

Is this a claims study? +
No. A claims study produces a marketing line from a one-time panel. We produce an ongoing responder read on your own customers, who responds, when, and why. You can build honest claims from it, but the deliverable is understanding, not a slogan.
Do you promise results, or a seal? +
Neither. We promise honest measurement. No guaranteed claim, no badge. A sourced, believable result earns more trust than a perfect-sounding one.
How is this different from testing a panel? +
A panel is matched strangers, measured once. We read your actual buyers, continuously, against each person's own baseline, so you learn about the people who pay you.
Is there a medical screener? +
In this configuration, no screener and no recommendation engine. It is a research and results-tracking tool. Where a physician is involved, they remain the decision-maker.
Does the companion coach my customers? Does that change the read? +
Yes, and we are transparent about it. The companion actively supports follow-through with reminders, encouragement, and your customer's own progress reflected back, not advice on what to take. Because of that support, the read reflects the real-world experience, your product plus the support around it, and we label it that way every time. It is what your customer actually lives, measured honestly.
What do founding partners lock in? +
Preferred terms, a seat shaping how the program is built, and first position in your category as the brand that measures honestly. Terms are set per partner.
What keeping more customers is worth

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.

Estimated added revenue
$18,000
a year, estimated from the numbers you set above
25
more customers kept, per month
$1,500
added revenue, per month
≈ 1.8x a first year of Rationally Well
The Founding Partner base offer is $4,000 (was $8,000, 50% off while early) and covers up to 300 active testers, then from $500 a month in the test phase for the measurement itself.
The pilot runs on up to 300 active testers, and the playbook it produces, who responds, who quietly quits, and when, applies to your whole base. That is what this models.
You set the improvement to model. Rationally Well measures and reports. It does not promise a result.
See your demo

See who is responding to your product.

Tell us about your brand and a good time to talk. We will build a responder snapshot on one of your products, on your own customers, and walk you through who responds, when, and what travels with the result, including the part that does not flatter.

Who is respondingExample
46%31%23%
Strong responders Partial No change
The kind of snapshot you would walk away with, on your own buyers, measured against each person's own baseline.
  • On a slice of your real customers, not a panel of strangers
  • The full picture, including the customers who did not respond
  • No spam, no sales pressure, no outcome promises