Orange could easily settle on a better market strategy by having blind rabbits do the picking than through the "market research" I was subjected to today. At least then, management would know they were betting the company on fuzzy luck done in the dark. Unlike the false sense of security they'll get when they receive their worse-than-random report from the undoubtedly over-priced research firm in an impressing binding and with colorful graphs.
Seriously, I had never imagined that a serious market study commissioned by a serious company in a serious business could be so... un-serious. But that was exactly what the questions and the procedure I was subjected to as interview volunteer today. They broke most, if not all of the, rules in the book on quantitative data gathering:
- Excessive amount of questions: They exceeded the recommended maximum range (20-30 questions) by more than 100%. A completely unjustified exception to standard interview practices given their trivial research content.
- Irrelevant precision: They started the questionary off with at least ten sets that require a complete ranking of seven options. I cared for no more than three or four options at the most, so forcing me to be specific about more than that brought only irritation and bogus answers.
- Extreme question bundling: An incredible thirty questions revolved around seemingly random bundles of four options put against four other options. You didn't get to pick and choose. It was either one or the other, which will led to completely unusable answers (my decision for picking option A over option B could be any of eight parameters).
- Puzzling options: "Which of these would you rather like? a) $500 b) $100 c) $1". They had at least five questions in this category and asked you to rank the obvious-from-the-outset alternatives.
- Almost no segmentation: Apparently, one size (almost) fits all at Orange. There was almost no segmentation questions on household income, number of text message usage, how many companies you've been with during the past year, etc.
Oh, and the questionary was completed on a PC from the eighties running a DOS-program that screamed a loud beep should you be so foolish to press the same button twice.
Basically, the study was a joke conducted with no sense of responsibility and Orange are fools if they prefer this data over our fuzzy blind rabbits.