mardi 6 janvier 2009

Marketing research and its discontent


Out-of-home advertising business is growing. To become a real business, OOH needs a measurement tool that advertisers and media buyers can trust.
Nielsen is building a pilot for a syndicated service, PRISM (Pioneering Research for an In-Store Metric). Wal-Mart took part in this pilot but will not be continuing. Probably because Wal-Mart does not share data with syndicators. Moreover, Wal-Mart is developing its own service with DS-IQ for its digital network. And it is too expensive. Sounds a little like Apollo ... (cf. post du 26 février 2008). The times do not lend themselves to spending large amounts of money in what is certainly uncertain research.

Out-of-home measurement is still prisoner to all the problems that revolve around panels. Difficult and close to impossible to recruit respondents representative of  shoppers, passers-by, visitors of all kinds in malls, retail outlets, museums, stadiums, etc.
OVAB in the USA sets  guidelines for Out-of-Home media, which are not different from what is expected from in-home media (esp. TV): Opportunity To See (and/ or Listen) OTS/OTL, contact (still a fuzzy concept), Time Spent Watching and Listening (a type of "stickiness", nicely called "Dwell Time").

How to recruit panelists for this kind of measurement? Nothing new: the usual quotas (geographic, demographic, psychographic), but all based on declaration instead of random surveys and observations. How far can we trust these declarations to be true enough to build representative samples and longitudinal panels? Can we trust them to produce explanatory variables (independant or controlled)? How many phone calls before finding a single panelist? How many intercepts for an interview? This is why such research is becoming more and more expensive; this is why there are more and more non-representative panels, catch-all panels...

Research, sociological as well as marketing, is uneasy: from panel recruitement, from low cooperation rates. Who agrees to participate in panels, to answer questionnaires? We need iconoclastic research on the research conditions. An epistemological approach to research conditions. Passive, non-intrusive research is in! Ethnographical research (Erwin Goffman as in "Frame Analysis", Harold Garfinkel as in "Cognitive Sociology", etc.) will bring the indispensible qualitative dimension.

Aucun commentaire: