Rather amazing news broke this week. The Market Research and Intelligence Association (MRIA) is shutting down – for good. As a firm that offers market research here in Canada, we at Tenato wanted to at least share a few thoughts on this development.
Although it is a shame to lose an industry association aimed at bettering professional standards, the demise of the MRIA did not surprise us. In fact, for the past several years we have been amazed time and time again at the abysmal ability of pollsters to predict various election results. A few years ago, we wrote a blog about the 2012 Alberta provincial election discussing possible reasons why the polls were so far off, and what alternatives existed.
If traditional polling techniques rely so heavily on people answering emails or using land lines, how can they possibly be accurate? Why is the media clinging so tightly to companies who use these methods when predicting election results?
Tenato went looking for alternatives, and didn’t need to look far.
In 2012, we had been testing a new online monitoring platform called Sysomos, and discovered that social media sentiment monitoring, along with tools such as Google Trends, seemed to more accurately reflect (and thus showed potential to predict) election results. We invited a member of Sysomos’ team, Jay Cloutier, to mine his election data to see if it could indeed be used this way; he wrote an interesting article which we posted here.
His data confirmed what we witnessed online: Sysomos’ social media monitoring beat polling for accuracy, hands down! If a candidate said something good, you’d see it on social media. If it was stupid, you’d see it on social media. As a candidate’s name is mentioned and discussed, you can see it in Google Trends. Why trust a weeks-old poll when online monitoring was up-to-the-minute, every moment? The numbers proved it.
It was then that Tenato realized that monitoring what people actually were saying (i.e., what they were already expressing on social media) was more predictive than asking them what they intended to do (i.e., polling). With such a striking difference in accuracy, we felt that traditional market research, especially in the form of polls, was a dying animal. Since that time, we have relied on many sources of online data, using them to cross-reference traditional methods, or even on a stand-alone basis. For example, if Google Trends says all keywords related to noodles (e.g., noodles, ramen, udon) are upward trending, it’s probably not a bad thing to be expanding your noodle business! Calling people and asking them if they are eating more or fewer noodles than last year? Tough to say.
The problem with many market research firms is that they specialize in only market research, and remain isolated from the rest of the world of marketing, often clinging to traditional academic methods. Meanwhile, the rest of the marketing industry is moving online and gaining a huge degree of scientific accuracy.
Many market researchers believe that market research should be separate from strategy (Tenato 2015 blog) missing the fact that advertising itself was becoming a form of market research (Tenato 2013 blog). The result? They miss the avalanche of big data (and hence, don’t learn how to mine it) and fail to realize that there are more efficient ways to predict human behaviour than making phone calls to land lines and asking people what they intend to do.
I know many in the industry will be sad to see the MRIA go, but I never felt that the MRIA had much to offer Tenato; the conferences, the topics… all seemed stuck in the past. While I do think the statistical principles enforced by members of the MRIA are important, it is a shame the MRIA did not lead the charge toward methods they could have been exploring. Ignoring data points is not what good market researchers should be doing.
Perhaps there is a future for a new kind of association one day; one that will educate students and professionals in the latest techniques to improve data accuracy and reliability for all those who need it.