I started my first blog end of March last year. As a market researcher I’m obsessed with figures. As a co founder of MarketResponse, a market research and consultancy in the Netherlands I discovered at that time that on the first of February next year, my company will celebrate its 25th birthday. Now as I started to write blog I found out two things. First of all that the countdown to the previously mentioned birthday was 666 days. Well, that is a message in itself isn’t it? Great coincidence or some bad guidance from hell? I did not know until I wrote down the birthday in figures (in Dutch annotation: 01022010. A Palindrome
Well that is something as well I thought and started to blog. The world we live in is full of hidden symbols and great numbers. As an experienced researcher I’ve presented numerous numbers. ‘Probably’ the number 1 was presented most. It seems that this was researched a long time ago. I will never forget the importance of probability distributions and statistical errors. In the early eighties I acted as project manager to measure the distribution of a leaflet called Shell Helpt. A consumer oriented leaflet, in which Shell gave numerous suggestions, dos and don’ts how to maintain your car successfully. The aim of the study was to measure the % of complete distribution on household level in the Netherlands (i.e. people in the sample had to report whether they had received a Shell Helpt leaflet, being the same species as we had sent them a few days before). Although it was a great study design with three stage sampling (random selection of cities, followed by random selection of streets, followed by quota sampling of 12 addresses) procedure I was not quite savvy to write a valid report. Statistical studies had indicated that in a certain street 5 consumers had to confirm ‘yes’ to be ‘sure’ that distribution had taken place. Still there was a small chance that people would say yes, when there was no distribution. On the other hand also there was a chance that respondents would report a ‘no’ whilst distribution actually had taken place, but we only had to find 2 no’s per street in order to know for sure. People tend to report more ‘yes’ more easily than ‘no’. I reported (after vast calculations) a 93% distribution success rate, added the standard error but multiplied it by two times a 1, 96 factor (used when looking for 95% confidence level). Obviously that was wrong. Client was not pleased after they found out and I felt bad. Yes it would cost the distributor money since I had been too generous. Actually I helped Shell, but not in the right way. Anyhow this is way before my career started to move. I still love statistics but probability sampling is something I leave to specialists. To me it is complexity. I will therefore report more qualitative experiences over the next few days.