Jerome Sopocko Jerome Sopocko founded Askia - software for surveys in 1994. Prior to that he was employed as a consultant by BMRB to design the TGI's PTK toolkit. Askia launched with askiaAnalyse a tool designed to facilitate the creation of crosstabs with specific emphasis on open-ended questions and multi-dimensional analysis techniques. Read the full biography here.

Man Up! The Trojans are at the Gate

a call to arms for techies

24th November, 2010

One of the most entertaining features of the Market Research industry is its fatalism. Each year brings a new contingent of Cassandras predicting the end of the MR world as we know it – and the Esomar September congress is usually prime time for the harvesting of such predictions. As I expected, this year we were again castigated for not ripping the whole thing up and starting again. Don’t get me wrong, pessimism requires intelligence whereas optimism doesn’t. And Troy was indeed destroyed in the end... so now we’re on the subject of horses, why the long face?

Well, our industry has indeed suffered tactical losses in our influence over the business world. First we lost the ‘Bores War’: Statisticians vs Accountants… Now management consultants are the natural advisers of the board of large companies instead of us. Then there are all the areas where we had the technical know-how but we let it slip through our fingers to other players; direct mail to geo-marketing specialists, web analysis gone to Google analytics, customer intelligence lost to CRM experts with SAP and social network analysis… well the jury’s still out on that but it’s likely to be a division of FaceBook Inc. Behaviour analysis is not collected by survey anymore but by Tesco cards. And when dunnhumby starts to seriously tap into Market Research, it won’t bother asking what we are (they already know that through the accurate digital traces we’ve left using our loyalty card) they’ll be able to concentrate on the “whys?” and the “what ifs?”

Those lost battles were no surprise to any assiduous MR conference attendee. I remember Laurent Flores talking about the dangers of disregarding CRM a good five years ago. We did (still do) try to compete in the online advertising analysis front but Google is simply too strong (and too good). And there’s Ray Poynter, of course, who’ll tell anyone that’s listening (and I am) that quantitative surveys will be extinct in 20 years – tragic news for me as the retirement age keeps on creeping up with no other career option than the lottery ticket I’m yet to buy.

We, as consumers, leave trails – foot and finger prints of every transaction we make and every interaction we engage in. Nothing really new: consumers have been leaving trails in our increasingly connected world years before the internet starting tracking them.

Swiping credit cards and walking through CCTV covered areas leaves traceable paths of our behaviour. And sometimes even a seemingly simple cash transaction at a cash register for the brand of our choice can leave a trail. Brands like Nespresso took 'belonging' to a new level by tracing coffee drinking behaviour, capsule by capsule, flavour by flavour through the use of an exclusive distribution channel. Point-of-purchase information at the simplest of retail establishments contributes to an increasing amount of highly granular information about consumer lifestyles. Whether it’s through automated pre-paid toll road systems (EasyPass/Fast Lane), mail-order subscriptions, or even charitable donations, one thing is for certain: we have always been, and will always be tracked and targeted as consumers.

But we live on. And I can but hope that market researchers still believe that a good decision is one that’s validated by a good old cross-tab; and why not? It does work.

So let’s look at some things we can still do something about: the two pressing challenges upon us are the dropping response rates and DIY surveys. The dropping response rate is not new… telephony response suffered badly from double-glazing salesmen who targeted the same people we did– how dare they! We thought we’d found the answer with online surveys and panel providers. However, panel response rates are now also falling dramatically, and we are looking round for something to blame.

Quickly the finger was pointed at long and complex surveys – especially those with conjoint matrices. “Haro sur le baudet!”- “Down with the donkey!” Our industry produces one Nobel Prize winner, Daniel McFadden, for his work on conjoint analysis and we are politely being asked to disregard it! I got into MR through mathematics; not only because it was beautiful (and it is) but because it works. I was changed forever by my first course on Factorial Analysis and when I discovered it could be used on raw open-ended data I had my epiphany. The second coming happened when I saw Steve Cohen present a paper about the launch of camera-dockable printers for Kodak. They managed to create a simulator predicting the market share (and return on investment) taking into account the pricing, the available features (all combinations of them) and the eventual presence of competitors. The simulator was so accurate that they used it for years after the study. We are talking about actionable MR here not just brand value and ad-recognition levels.

Let’s not drive our clients into the hands of the DIY-ers by implying it is OK to dumb down surveys. The solution to low response levels is not to simplify surveys. It’s not about asking for gender, age, social grade and would we recommend the product to our friends. By doing so, we will have let the wolf into the sheep house and the cat among the pigeons. Our clients finally realised that asking a few questions on a web form and cross-tabbing the results was no rocket science. We tell them that DIY research is not a good idea… why? Because we know better; we know how to sample, we know how to word questions; we’ve been doing it for 70 years. Please, if you do con some cash out of someone, it does not help to remind them for how long you have been doing so.

Look at it from the point of view of the big corporations: they have the panel (their customers); access to communities and basic DIY software is free; and any one member of their marketing department can write a questionnaire (as long as it doesn’t have any of that difficult technical stuff). It’s about time we reminded them about the value we can add to their project!

People like to give their opinion; fact. But they do need to feel that you listen. In other words, they want an intelligent conversation. A good online survey needs to be informal, pick on relevant answers and enquire intelligently about them. Opinionway in France has developed a simple algorithm to dig into open-ended responses and ask specific questions depending on the presence of topical keywords. Doron Meyassed from Promise, in the August issue of Research, declared with a clear dose of provocation that we should reject respondents with an IQ below 120. I disagree, mostly for representativeness and also because I like to fill in surveys. However, how about we treat our respondents like they all had an IQ above 120? We need to ask them hard, intelligent questions. And choosing one product among others – with a wide set of features and prices - is a hard, intelligent question. And yes it does help if the layout is nice and the features are visible interactively in a bitmap rather than text based.

They want to be entertained. BrainJuicer, alas not a client, has included in their Esomar survey a trading game. I’d like to know about their response rate. I ALWAYS fill in a survey from BrainJuicer. I know I will be challenged, excited, amused, interested and not patronised. Incidentally, they develop their collection software in- house and do not use an off the shelf package. We need to make sure that our scripts are written using Flash, Silverlight or HTML 5.0. Surveys need to be fun and that does not mean including sliders with a smiley face. Also other technologies are there to be used: we can put respondents in a more “real life” environment, pick products from a virtual shelf, measure in real time their feelings towards a TV ad, understand text in natural language, record their voice, record their faces.

Finally, people want to belong. We need to give them feedback at the end of the survey. A classification of their responses, some basic live results, a relevant redirection, some trivia about where they come from, a word cloud linking to the responses of an open-ended question – or better, a forum, a focus group, some interactive game where they can interact with other respondents.

At the last ASC conference in April 2010 MR software providers were accused of not having embraced the social media revolution. Maybe we are another victim of bad research; we only deliver what our clients expect us to do. Get in touch with your software providers. We’re geeks but we will listen because we also need to get some excitement out of this industry. Time to think outside the tick-box, I say.

Jerome Sopocko

Comments on this article

Want to share your thoughts...?

Want to share your thoughts?

NOTE: Please note that this board is moderated, and comments are published at the discretion of the site owner.