01 · Untrusted sources
Who answered?
Opt-in panels leak bots, AI agents and professional fakers, and most of it gets past the usual checks.
99.8%
of attention checks passed by an AI agent built to take surveys
Westwood, PNAS, 2025
Independent brand lift studies
Brand Lift that can’t be spun.
SpinProof proves whether your advertising changed what people think, with verified people, a matched control and a plan locked before fieldwork. Every number comes with the evidence behind it.
- Verified people, never synthetic
- Plan locked before fieldwork
- First read in week one
- Evidence File in every study
The problem
Brand lift is a number almost nobody can check.
Budgets still move on these numbers. Three things get in the way of the truth.
02 · Unproven exposure
Who saw the ad?
Exposure is assumed from where people were surveyed, and the company that sold the media often runs the test.
41%
of marketers name walled-garden reporting limits as a top barrier to accurate measurement
TransUnion and EMARKETER, 2025
03 · Unevidenced results
Can anyone check it?
Headlines arrive without the plan, the raw data or a trail from each number back to the answers.
0 of 3
recent brand lift reports we reviewed proved their headline on the full sample
SpinProof review, anonymised
21–38%
fraud found across four large panels and two panel exchanges
Rep Data study, 2025
~30%
of survey fraud caught by traditional data cleaning
Rep Data, 2025
12%
of opt-in respondents under 30 said they could operate a nuclear submarine
Pew Research Center, 2024
What spin looks like
Same study, two stories.
A real report we reviewed led with ‘+29% purchase intent’. It came from one age band: 39 exposed people against 32. On the full sample, the honest read was no measurable change.
How it was reported
+29%
Purchase intent, age 25 to 34 only. 39 vs 32 people, p = 0.20.
How SpinProof reports it
+3.1
Points on the full sample, 95% range −6.7 to +12.9. No measurable change.
Introducing SpinProof
Brand lift studies that come with their own proof.
Three studies on one method. Every one uses verified people and a matched control, locks its plan before fieldwork, and ships with the Evidence File.
Creative Test
Will this ad move people before we buy media?
300 verified people across exposed and control cells, randomised inside the survey, before the media buy.
From £2,500 per creative
Campaign Uplift
Did the campaign shift perceptions in market?
600 verified people, proven exposure against a matched control, and a live read from week one.
From £4,950 per study in a shared survey wave
Brand Tracker
Is the brand building between campaigns?
300 verified people a month, plus share of search, competitor ad activity and AI answer visibility.
From £1,750 a month per brand and market
How it works
Five steps, three human gates, one evidence trail.
01
Lock the plan
KPI, sample, sub-groups and analysis plan, hashed before fieldwork.
02
Recruit real people
Meta, QR codes in the media, publisher lists and panels, all through fraud and attention checks.
03
Prove exposure
Exposed against a matched control, confirmed by recognition, a decoy ad and media data.
04
Agents run it, people sign it
Ten agents do the work. Researchers sign three gates. Code does the statistics.
05
Release with evidence
Live portal, a page-one answer, raw data and the Evidence File.
The client portal
Results you can see, check and act on.
Every SpinProof client gets a live portal instead of a 30-page PDF. Click through the views below, or open the full demo.
Page one answers four questions.
Did it work, how sure are we, why did it work and what next. A verdict on the KPI fixed before fieldwork, with its 95% range, the sample behind it and a costed recommendation.
Open the demo portalDemo portal: fictional airlines, illustrative data.
A first read in week one.
The running estimate and its range update as verified answers arrive, with quotas by channel. The plan locked at Gate 2 decides the final answer, not the day it looks best.
Open the demo portalDemo portal: fictional airlines, illustrative data.
When nothing moved, we say so.
A sale that lifted ad recall but not consideration is reported as no measurable change, with what to do about it. No hunting for a sub-group that looks better.
Open the demo portalDemo portal: fictional airlines, illustrative data.
What the money actually bought.
Persons lifted, cost per person lifted, the frequency curve, and winners by placement and creative, so a finance director can act on the result.
Open the demo portalDemo portal: fictional airlines, illustrative data.
Every respondent accounted for.
Sample flow by channel, every exclusion with its reason, and the false-recall rate on a decoy ad that never ran.
Open the demo portalDemo portal: fictional airlines, illustrative data.
Locked, signed and traceable.
The plan hashed before fieldwork, the three human gates, the agent run log and a trace from every reported number to its source rows.
Open the demo portalDemo portal: fictional airlines, illustrative data.
The Evidence File
Every number traces back to a raw answer.
The audit trail other studies leave out. It ships with every SpinProof study, with the raw data and crosstabs. Always included, never on request.
- Method lockThe plan, hashed before fieldwork
- Sample flowEvery invite and exclusion, with its reason
- QualityFraud, attention and decoy-ad checks
- Agent run logEvery action, and the model or code that ran it
- Sign-offsWho approved each gate, and when
- Results traceEach number linked to its source rows
Why SpinProof
Independent, cross-channel proof, priced for the mid-market.
Independent
We don’t sell media, so we never mark our own test.
Any channel, offline included
Print, outdoor, venues, audio, events and retail media, where platform lift tools can’t see.
Live from week one
A first read while the campaign can still change, not a PDF after it ends.
From £2,500 a study
Managed brand lift from Nielsen and Kantar lists at $7,500 to $23,500, plus sample.
Industries
Built for categories where the brand decides the sale.
Airlines
Prove the route launch worked, not just that the ads ran.
Independent brand lift for airlines: prove whether route launches, cabin campaigns and sales moved consideration against your competitors, with verified flyers and an Evidence File.
Brand lift for airlinesRetail
Retail media is measured by the people who sell it. SpinProof isn’t.
Independent brand lift for retailers and brands buying retail media: measure what campaigns changed in shoppers’ minds, separately from the network that sold the ads.
Brand lift for retailLearn
Guides for brands who want the truth.
ArticleBasics · 6 min read
What is brand lift? A plain-English guide for brands
Brand lift explained: how it is measured, the metrics that matter, absolute vs relative lift, and what makes a brand lift result worth trusting.
Read the articleHow-to guideMeasurement · 7 min read
How to read a brand lift report in ten minutes
An eight-step checklist for reading any brand lift report: the KPI, the sample, the full-sample result, absolute lift, ranges, exposure, exclusions and data.
Read the guideArticleData quality · 6 min read
AI bots are taking your brand surveys. Here’s what that means
AI agents now pass 99.8% of survey attention checks and panel fraud runs at 21% to 38%. What that means for brand research, and how to protect it.
Read the articleFAQ
Questions brands ask.
What is SpinProof?
SpinProof runs independent brand lift studies for brands, agencies and media owners. Each study measures whether advertising changed what people think about a brand, by comparing verified people who saw the campaign with a matched control group who didn’t. The plan is locked before fieldwork and every study ships with an Evidence File.
What is brand lift?
Brand lift is the difference in a brand metric, such as awareness, ad recall, consideration, favourability or purchase intent, between people who saw a campaign and comparable people who didn’t. SpinProof reports it in percentage points with a 95% range, so you can see both the size of the effect and how sure the study is.
What does “Brand Lift that can’t be spun” mean?
The primary KPI, sample size, sub-groups and analysis plan are fixed and hashed before fieldwork starts. Deterministic code runs the statistics, a researcher signs the report, and every reported number is traced to the raw answers behind it. Nobody, including us, can pick a flattering sub-group or change the question after seeing the data.
What is the Evidence File?
The Evidence File is the audit trail that comes with every SpinProof study: the method lock, the sample flow, the quality checks, the agent run log, the gate sign-offs and a results trace linking each number to its source rows. It ships with the raw data and crosstabs, so a client, their agency or an auditor can check any figure without asking us.
How do you make sure respondents are real people?
We recruit through channels tied to the media and to consent, such as QR codes in the media, publisher lists, Meta ads into a screener and consented panels. Every respondent passes fraud, attention and duplicate checks, and a decoy ad that never ran measures over-claiming. Every exclusion is logged with its reason and shown in the Evidence File.
How do you prove that people in the exposed group saw the ad?
Exposure is confirmed three ways: recognition questions about the real creative, a decoy ad that never ran, and the client’s media data on placements, timing and frequency. Where possible we also recruit from the media itself, for example with QR codes at the venue, so exposure is evidenced rather than assumed.
Which channels can SpinProof measure?
Any channel, including the ones platform lift tools can’t see: TV and online video, social, display, print, outdoor, venues and transport, audio, events and retail media, on their own or as a mixed campaign.
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.