SINGLE SOURCE OR FUSION – AN ATLANTIC DEVIDE?, Written by Manfred Mareck, Research Marketing Ltd
Evaluating multi-media campaigns requires a detailed knowledge of the overlap between media, for example Print and TV. To get this data media planners ideally need single source surveys that capture all media usage in one study. The alternative is to fuse or integrate individual media surveys into a ‘single’ source.
Many Techniques – no Real Winner
There are many ways how statisticians can fuse, merge or integrate the data of two or more different studies. Different companies, such as Kantar, Telmar, RSMB, Czaia Marktforschung and many others have developed various techniques. The statistics behind are quite complex and not all fusion projects have managed to preserve the integrity of the original pre-fusion data. This is a particularly important issue when integrating different media surveys. Media owners don’t like to see their net reach change in the fusion process, especially if the data goes south. Rumour has it that something along these lines happened with Project Unity in Canada, which attempted to fuse the different media currencies for Newspapers, Magazines, TV, Radio and Outdoor. As far as I can make out the fusion has taken place (twice actually), but nothing much has been published. All it probably needs is one unhappy media owner who can claim the copyright to the original data and legal problems raise their head. Indeed, data integration is generally less of a technical issue, as Steve Wilcox of RSMB pointed out at a recent meeting of he UK Media Research Group. There are no right or wrong fusion techniques, they are either appropriate for a specific task or not. In other words there’s little point in trying to develop a ‘one size fits all’ fusion gold standard.
Two other interesting, large-scale projects are currently under way on both sides of the Atlantic: Project Apollo in the US and Touchpoints in the UK.
Encouraged by Proctor & Gamble, Arbitron and Nielsen Media Research announced earlier this year their plan to set up a single source super-panel of US consumers. This panel would provide data on all media exposure combined with an actual measure of shopping behaviour. Various data collection methods are deployed to get the information. Television and Radio will be measured through Arbitron’s portable people meters (which incidentally can also pick-up signals from in-store radio); online surveys will collect Print, Internet and lifestyle data, Nielsen’s HomeScan will track purchase behaviour. The set-up is currently being piloted and, if successful, the aim is to sest-up a 30,000 strong household panel representing close to 70,000 individuals. The final data will provide insight into media usage, life style and purchasing behaviour of the American public. The promotional literature gives various scenarios; it mentions for example the mother, who hears a radio commercial for McDonalds at 07:30 in the morning and goes to have lunch with her daughter at McDonalds at 13:00, thus demonstrating the impact of advertising. I did say this comes from promotional literature!
Touchpoints in the UK uses a different route and is based on data fusion of the main media currencies. But rather than fusing one currency with another (for example Print and TV) followed by fusing a third currency onto the newly fused database, Touchpoints first creates a separate Hub Survey, using an established TNS access panel. This central survey is modelled along the lines of the BBC’s time budget study called Daily Life, where respondents use a PDA device to keep a weekly digital diary to record their activities, including media usage. Touchpoints uses the same methodology combined with a self-completion questionnaire to collect additional in-depth data on media preferences and various other qualitative issues. The Hub survey, itself a rich source of data in its own right, will provide the hooks (in form of media use by time) for the fusion of the official radio, TV, print and outdoor currencies without changing the original results from these individual surveys. At least theoretically no media owner will have cause to prevent publication of the data.
Another key difference to Project Apollo is that Touchpoints is way past its piloting stage. Fieldwork for the Hub Survey has been completed and RSMB will now run the actual fusion process.
What’s interesting is that none of these projects have been initiated by any individual media owner. Project Unity is the brainchild of the Canadian Media Directors’ Council, Procter & Gamble encouraged two major US research companies to develop Apollo and Touchpoints is an initiative of the IPA (the trade body of UK advertising agencies). If they succeed, true multi-media evaluation will take a big step forward.
Research Marketing Ltd
12 Alexandra Road
London W4 1AX
T: +44 (0)20 8995 7057