Side by side
| Criterion | Rascasse | GWI |
|---|---|---|
| Method | AI-based analysis of aggregated digital behaviour signals (search, social, video, events, sponsorships, points of interest), not surveys | Online survey research |
| Coverage | 500,000+ data points in 200+ countries, including 50,000+ brands and 100,000+ media channels | 2 million+ interviews annually, 50+ markets, 35 billion datapoints annually |
| Update frequency | Quarterly | 4 waves per year |
| Historical depth | Up to ten years (4 to 10+ years depending on the market) | 15+ years of trended insight |
| Finest geographic level | Region, city and postal code, where available | Market and region |
| New audiences | Any audience you define, live in 24 hours; a new brand in 2 hours | Brands and categories in the questionnaire |
| Measures attitudes | No. Behaviour only | Yes: lifestyle, values, motivations |
| Access | Web (Rascasse Solution), exports and decks; API and MCP as part of your agreement | Web, API, integrations |
Where GWI is stronger
- Attitude and lifestyle depth. Values, motivations and self-reported lifestyle are the core of the offering, and they cannot be derived from behaviour.
- Longer trended history. More than fifteen years of consistently measured data, against our four to ten years. It cannot be rebuilt afterwards.
- A stable instrument. The same questionnaire across waves and markets makes some comparisons clean in a way behaviour coverage, which varies by market, does not.
- Segments by attitude. Segments can be built on stated values. Behaviour data can only approximate this through interests.
Where Rascasse is stronger
- Geographic resolution. Region, city and postal code, where available, against market and region. Site selection, out-of-home and regional tests all happen below the level a survey sample supports.
- Audiences you define. Any audience from brands, clubs, artists, events or media, live in 24 hours. For partner search this is decisive: the interesting property is often one nobody would have put into a questionnaire.
- Breadth of what is measured. Any brand, club, artist, event or medium with observable activity can be part of an audience, not only the ones that made it into a questionnaire.
- One method in 200+ countries. The same approach everywhere, including markets outside a survey programme.
- Inside your AI assistant. As part of your agreement, via API or MCP from Claude or ChatGPT.
Cadence is not the difference
GWI publishes four waves a year, and Rascasse updates quarterly. On cadence the two are comparable, so choose on something else. What differs is the unit of measurement (an answer against an observed behaviour), the geographic depth, and who decides what is measured: the questionnaire designer, or you, when you need it.
When to use both
These two combine cleanly because they barely overlap. GWI describes what kind of people a segment contains and what they say motivates them. Behaviour data shows where those people are and which properties, brands and events they engage with disproportionately. A segment defined by attitude in GWI and located by behaviour data is a stronger brief than either alone. See segmentation and personas and psychographic segmentation.
Making the case to finance
Whichever provider you choose, the budget owner will ask what it costs, what it delivers and how to check it. With Rascasse the Annual fee per market or region is agreed upfront, own audiences come as credits (one credit is one audience in one country), and every result is a chart, a deck or an export that someone outside the analysis can check. Teams without their own analysts can buy the answer instead, as an Insights Report.
How to test it
The fair test is one real question, answered both ways. Pick a question you already answer with GWI, for example where a GWI segment concentrates by region and city, or which brands and media it over-indexes on. In a Proof Case we answer it for one market and up to three audiences in three weeks, at a fixed price, and present the result to the person who decides on the budget. You then put both answers next to each other and judge where the difference matters for your decision.
Sources
- GWI: https://www.gwi.com/, retrieved 13 September 2026.
- Rascasse figures: our own published coverage, October 2026.
We do not measure or test other providers' products. Everything stated about them here is their own published claim or description. If a figure has changed, tell us and we will correct it.
