𧢠Coach T's Recap
Clean data means confident decisions. We vet every panel provider before we use it, check every response with our providersâ screening and our own checks, then clean and weight the data to census population counts before it lands in your dashboard.
Cleaning the data means checking every response before it counts. Hereâs what we check, who checks it, and what happens to a response that doesnât pass.
Before the survey
We vet every panel provider before we use it, testing it on quality, consistency and how many people it can reliably reach. After that, we keep monitoring how it performs.
During the survey: two layers of checks
Provider checks. Panel providers screen their own respondents. These typically include device and IP anomaly checks, geo-location clues, digital fingerprinting and reputation-based screening.
Our checks. We run an in-house engagement check, and AI-powered quality scoring that looks at 175+ behavioural, technical and response-based signals, including speed, answer patterns and the quality of written answers.
After the survey
We clean the data and weight it to census population counts. Where we use several panel providers, we also adjust brand funnel metrics for known differences between panels.
FAQs
Who checks the responses?
Every response is checked by the panel provider that recruited the respondent, and by us. Each providerâs screening differs, and we apply our own checks on top.
What happens to a response that fails a check?
Itâs removed before analysis, so it never counts towards your sample size or reaches your dashboard.
How do you vet a new panel provider?
Before we use a provider, we test it on quality, consistency and how many people it can reliably reach. After that, we keep monitoring how it performs.
Why do you weight the data?
Quotas on age, gender and region (and ethnicity in the US) get our sample close to your market. Thatâs what we mean by nationally representative, or ânatrepâ. Weighting to census population counts then closes the remaining gap, so the sample reflects the population.
Will cleaning change my trend?
The same cleaning rules apply across a market, so your brand funnel results stay comparable. Where we use several panel providers, we adjust for differences between panels, so a change in the mix doesnât show up as movement in your brand funnel. We flag statistically significant changes in your dashboard with green and red arrows, so you can tell real movement from normal variation. What is statistical significance?