How it works
Locked before fieldwork, signed at every gate.
Agents do the repetitive work, deterministic code does the statistics, and a researcher signs off three times. You get the result and the evidence together.
The short answer
A SpinProof study locks its plan before fieldwork, recruits verified people through the channels they already use, proves who saw the campaign, compares them with a matched control, and releases the result with an Evidence File. Agents do the repetitive work, deterministic code does the statistics and a researcher signs three gates.
Five steps.
- 1Lock the plan. The primary KPI, sample size, sub-groups and analysis plan are agreed with you, then hashed and timestamped before the first invite goes out.
- 2Recruit real people. Respondents come in through Meta ads into a screener, QR codes in or beside the media, publisher and partner lists, panel top-ups and SpinProof’s own opted-in panel. Every one passes fraud, attention and duplicate checks.
- 3Prove exposure. The exposed group is confirmed by recognition of the real creative, a decoy ad that never ran and your media data. The control group is matched people from the same audience and window who weren’t exposed.
- 4Agents run it, people sign it. Ten agents draft, recruit, check and write. Deterministic code runs the locked statistics. A researcher signs each of three gates.
- 5Release with evidence. You get a live portal, a page-one answer, the full results, raw data, crosstabs and the Evidence File.
Three human gates.
- Gate 1: a researcher approves the study spec and the survey before anything goes out.
- Gate 2: the analysis plan is locked and hashed before fieldwork opens, so nobody can choose a flattering sub-group after seeing the data.
- Gate 3: a researcher checks the report against the Evidence File and signs it before release.
Three study designs.
| Design | Exposed group | Control group | Best for |
|---|---|---|---|
| Creative Test | Target audience, randomly shown the ad inside the survey | Same pool, randomly shown a different ad or none | Creative decisions before the media buy |
| Campaign Uplift | People who saw the campaign in market, confirmed by recognition, a decoy ad and media data | Matched people from the same window who didn’t see it, or a pre-campaign wave | Proving a campaign worked |
| Brand Tracker | Monthly waves of the target audience | Pre-campaign baseline and competitor brands | Brand health between campaigns |
How big a study needs to be.
Every study states the smallest lift it can reliably detect before fieldwork starts. Our standard is 300 verified people per cell. The table assumes 95% confidence, 80% power and a 30% baseline.
| People per cell | Smallest detectable lift |
|---|---|
| 150 | about 15 points |
| 200 | about 13 points |
| 300 (SpinProof standard) | about 10.5 points |
| 400 | about 9 points |
Where respondents come from.
| Channel | What it does | Watch-out |
|---|---|---|
| Meta: Facebook and Instagram | Paid ads into a screener, targeted by location, age and interests | Targeting can widen, so the screener does the real qualifying |
| QR codes in print, outdoor and events | Invites in or beside the media, which also evidences exposure | Skews to engaged readers, so it is balanced with other sources |
| Publisher and partner lists | Invites a publisher sends to its own audience | The publisher sends as data controller, with wording agreed first |
| Panel exchange | Programmatic top-ups to fill quotas fast | Cost per complete rises as the audience narrows |
| SpinProof panel | Opted-in past respondents, re-contacted by WhatsApp or email | Contact frequency is capped to avoid professional respondents |
Ten agents, one job each.
| Agent | Job |
|---|---|
| Intake | Turns a booking into a study spec |
| Intelligence | Builds the context brief from competitor ads, share of search, AI answer visibility and news |
| Questionnaire | Builds the survey from locked templates, with the decoy ad and attention checks |
| Channels | Runs recruitment across every channel and fills quota cells |
| Verification | Screens fraud, bots, speeders, duplicates and false exposure claims |
| Fieldwork | Runs invites, quotas and incentives |
| Analysis | Runs the locked statistics in deterministic code |
| Reporting | Writes the report from computed tables and checks every chart against them |
| Evidence | Assembles the Evidence File |
| Norms | Adds each finished study to the benchmarks |
Six rules that keep it independent and lawful.
- Real people only. Reported results come from verified humans. Synthetic answers are only ever used to pre-test wording, and are labelled.
- Research, not marketing. No sales message goes in any invite or survey. The ICO treats marketing dressed as research as direct marketing.
- Opt-in first. WhatsApp needs opt-in before any business message, and privacy notices name every channel used.
- Clean incentives. Cash or neutral vouchers only. The MRS Code bars a client’s own goods or services as research incentives.
- Locked analysis. Method, KPI and sub-groups are fixed before fieldwork. Nobody edits results, and flat results are reported.
- UK GDPR by design. Consent, data minimisation, retention limits and processor agreements across every channel.
Questions
About the method.
Who decides the KPI?
You do, with us. We agree one primary KPI before fieldwork, such as consideration, and lock it at Gate 2. Other metrics are reported too, but the headline is always the KPI that was fixed in advance.
Do agents answer the surveys?
Never. Agents handle operations such as drafting, recruiting, checking and assembling the Evidence File. Every reported answer comes from a verified person, and the statistics run in deterministic code.
What if the sample doesn’t fill?
The Channels agent tops up whichever cell is behind from other sources, and the portal shows quotas by channel every day. If a hard audience can’t be filled at the agreed price, we tell you before fieldwork closes, not after.
Next step
Run a pilot study on your next campaign.
Pick a live or upcoming campaign in any channel. We agree the KPI with you, lock the plan before fieldwork, and you get a first read in week one, then the full report with its Evidence File.