Centralized vs. Decentralized Research: The Verdict Is In
Posted by Jeffrey Henning on Wed, Jun 30, 2010
My "Tweet Off" compatriots Kathryn Korostoff of Research Rockstar (@ResearchRocks) and Cathy Harrison (@VirtualMR) of Chadwick Martin Bailey are bickering again, and Cathy has invited me to issue a verdict on their argument:
- In Market Research Decentralization: Power to the People, Kathryn argues that the biggest problem facing market researchers today is that their research findings do not inspire action. By involving coworkers in the research process, from start to finish, letting them even do much of the research, they will be more likely to take action on the results. As decentralization occurs, market research departments need to rethink their goals: they should provide resources to coworkers doing research (e.g., "standardized questionnaire templates, sample sources, in-house panels and communities"), provide training, create and enforce policies for such research, while continuing to manage complex studies centrally.
- In Be Wary of Decentralizing the Corporate Market Research Function, Cathy's concern is not that action isn't being taken but that decentralized research will inspire the wrong actions. She says that market research seems simple enough for coworkers to take on themselves, but only because of "not knowing what they don't know". This anosognosia can lead to poorly designed surveys that lead organizations to decisions that damage the brand, the relationship with customers and corporate profitability. Centralized research departments serve as "a hub of dedicated professionals who work closely together and share information", providing "a holistic view of their target audience and building synergies across projects".
Well, Cathy, I'm old enough to have seen this movie before. I'm sure the folks in the typing pool who had spent years learning touch typing and had high words per minute thought their department would always be needed, especially after everyone was upgraded from typewriters to Wang word processor systems. I fondly recall the graphic arts department, whose staff certainly produced more attractive looking presentations than I do without them. I remember the MIS department complaining as executives and staff brought in IBM Personal Computers behind their backs. And the finance department was none too pleased when executives began doing their own modeling with VisiCalc. The market research department that wants to remain relevant should follow some of the nuanced advice that Kathryn offers.
Because the irreversible long-term trend is towards providing employees the software to help them better do their own jobs: from typing, to writing presentations, to crunching numbers, to doing surveys.
Do you think you have a research specialty so complex that it is immune to this? For instance, say you have industry benchmarking data and sophisticated modeling algorithms that put the potential of the retail sales opportunity for new products into perspective. Well, sorry, but Skuuber is an automated system that packages just that expertise for $4500: far cheaper than a custom consulting study. If you thought Do-It-Yourself surveys were bad for researchers, wait until you see the growing number of narrow research applications that will automate those tasks you thought required your expertise.
In 1996, Perseus Development was two people: my business partner Rich Nadler and myself, providing CAPI (Computer Assisted Personal Interviewing) software to market research firms and departments. When I told Rich that eventually web surveys would be as ubiquitous as presentation software, he thought I was crazy. A year later he had warmed to the idea, and we released SurveySolutions for the Web. We rode the survey decentralization trend to 50 employees before we sold the company to Austin Ventures in 2006.
Sorry, Cathy, the verdict is in, but I didn't deliver it. End users did. Today, more surveys are done outside market research departments and market research firms than inside them. The days of the traditional market research department are numbered; they just haven't been counting them down.