Head to head
Cleared- Win share
- 65%
- 95% interval
- 56% – 73%
- Your bar
- 50%
- Dough default
- 50%
78 wins, 42 losses, 8 neither. “Neither” is shown and not counted as a win. Timing-flagged battles are excluded.
Report
Example report.
Simulated. Not a client study. The sections and the verdict rule are the ones a brand receives.
Concept test · New York
Keel Lemon Water. New label against the current bottle.
VerdictNot cleared
The new label won the pick. It missed the price. One missed bar and the study does not clear.
Bars
Price
Not cleared- Conservative share
- 38%
- 95% interval
- 30% – 47%
- Your bar
- 50%
- Dough default
- 50%
Anchor $2.49. Generous share 52%. The bar clears only if the conservative count clears it. This is what people say they would pay — not a sales forecast.
Liking
Not tested- Status
- Left off
You left this bar off. Had it been on, it would be the share who said love it or like it, asked before any choice.
First look
Each pack, rated on its own, before the battle.
New label
n = 120 · top box 33% · top two 68%
- Love it
- Like it
- It's okay
- Don't like it
- Really don't like it
Current pack · benchmark
n = 120 · top box 23% · top two 53%
- Love it
- Like it
- It's okay
- Don't like it
- Really don't like it
What mattered
Shoppers mark which pack attributes matter most and least. The ranking is a guide to the next design — not a second verdict.
- 1Looks cold on the shelf
- 2Easy to read the name
- 3Feels premium
- 4A bottle I already know
- 5Looks like it is trying
Stated, beside chosen
After the choices, shoppers rank the set. Mean pair agreement 64%. Top-pick agreement 71%(112 comparable). Speech and choice are both kept. They are allowed to disagree.
Price demand
What they wrote
“The new one looks colder. I'd take it if it cost the same.”
“I know the old bottle. This one feels like it is trying.”
“I'd buy the new label at the bodega. Not at that price for the week.”
Emails, links, handles and phone numbers are removed before brands see answers; names are not detected.
What to do with a “not cleared”
The pick says the label can win attention. The price says $2.49 is ahead of what these shoppers will grant it. The next study is a price, or a label that keeps the cold look and spends less effort announcing itself. The report will say if that one clears. It will use the same words.
Method, as written in the report
- Confidence
- 95%, fixed by Dough; brands set the bars, never the confidence.
- Result rule
- Cleared when the whole interval sits at or above the bar, not cleared when it sits below, too close to call when it straddles; fewer than 10 answers reads as not enough responses.
- Head to head
- Each respondent's single direct battle between the design and the benchmark (scoring rounds, timing-flagged battles excluded); Wilson interval. Neither answers are shown and not counted as wins.
- Liking
- Share answering Love it or Like it, asked before any battle. Against your current pack it is the paired difference on the same respondents.
- Price
- Stated willingness to pay at or above your price, never a sales forecast. Price bands are ranges, so the share is a range: it clears only if even the conservative count clears the bar.
See this for your own design.
A person reviews every request and writes back.