Friday 29 January 2010

Lies dammed lies and statistical bias



There are many reasons for wishing to pursue this ongoing theme which could last about three years, so I'm in no hurry. Also we've had sufficient time after the last Google Trends chart embed to find out that it isn't a dynamic script which would have caused extra work as I want those charts to be fixed in a point of time, and not self updating.


I'm also always looking for different interpretations in the comments from people who like this sort of thing or even blunt finger pointing at glaring examples of oversight or bungled conclusion so do get stuck in as this one is more journey than destination. But in principle it's a monologue hopefully interspersed with bits of conversation about how to sensibly approach the world of quantitative research and analysis. I don't think it's any secret that I'm more interested as a planner in brand DNA issues. I'd rather find out more about distribution and assembly matters than a TGI analysis. However that doesn't mean I ignore customer segmentation data. I just happen to think that most is not only badly conceived, it's then poorly executed and to compound matters it's given the sort of analysis that demeans all the work put into it. I also see very little meanings discussion in meetings with clients and even less externally on blogs. That may be because it's boring but even just articulating some of the thoughts and conclusions I've encountered is useful and for you too I hope.


One example I'd like to give is when working with NOKIA who really do have an awful lot of love from me for reasons they dont like to talk about (durability) but nevertheless around the time I was in orbit to a fair chunk of Helsinki's GDP the global market share for NOKIA was around 40 percent. Profits were dizzy and breathless, the manufacturing plant in Beijing looked like the most progressive work environment in the People's Republic of China, R&D in Oulu, Finland was staffed by the kind of people who weren't just boffins with a surplus of cash, there were even a few cool people paid to be imaginative. I blogged about it over here.


But it seemed to me at the time that NOKIA were on the precipice. Difficult to explain but it felt like there was little upward trajectory left ahead. There simply wasn't that much more room to grow given their handset segmentation strategy. Too many phones, too many segments and a notable absence of wow or excitement. The iPhone had just come out (mine was conveniently stolen at the IPA awards about 24 hours after I bought it) and the market for apps hadn't matured sufficiently so the future wasn't as fertile as it now looks. The single largest potential for increased market share was the rapid decline of MOTOROLA for reasons I wrote about over here. Other than that Palm and Blackberry were cranking up their game and a number of the Asian competitors were increasingly improving their products. That's HTC (Now the Google Nexus manufacturer), ASUS, Samsung, LG and all the other Asian posse, who do mobile marketing hygiene really well.


Having the conversation about how to defend market share rather than retail obsessed new product launches was not on the cards with NOKIA. To be fair, by the time you''re in a meeting discussing the launch of a model that looks suspiciously chunky despite it's geo utility specs they're just into selling the product so it doesn't gather dust on shelves and so macro discussions of how to spend the marketing budget more wisely aren't on the table. We have however since seen NOKIA's profits tumble and it's difficult to see how they can resecure their formidable market advantage again. A situation that was pointed to by the statistics long before they got there and mainly on a common sense hunch. I'm reminded that at that time I found their purchase of the mapping company Navteq for 8 billion US Dollars as really interesting and yet we now see that it's given away for free to match up against Google's mapping offer - That's a lot of money turning on a dime. This is further evidence that strategic planning is increasingly diminished in a fast moving world.


And it's on a hunch that I want to go back to the topic of 2012 as a search term. My assertion (completely my own and thus riddled with uncertainty, error and naturally intellectual hubris) is that the 2012 search term will become a byword for a sort of post pre millennial tension not unlike the sort we experienced from the Y2K bug. I'd completely forgotten how angst ridden a lot of people were at the time but the 2012 search term works well on a similar level because it's a numbers based term and thus transnational, which is really important. And also it's a sort of third party projection for every conceivable worry over the next three years or so. I've started rereading my Black Swan again and Taleb reminds us that in the Pleistocene period, the sort of randomness we experience on an almost daily basis through sheer human activity simply didn't occur. There's a hint of the singularity  that emerges here too, but in short I think the ride get's wilder as humanity progresses.


You can quote me on that.


To paraphrase my new favourite Irishman, Terrence McKenna "we're caught in a white knuckle race between education and extinction" so I assume the 2012 search term will serve as an indicator for quantitative analysis of the unquantifiable. That sounds about right doesn't it? However as I've blathered on a fair bit I'm going to return to the topic afresh and will only embed a fresh update of the previous chart which shouldn't have any significant changes. I'll be taking a look at country breakdowns of 2012 search terms. It's not only interesting, it's hugely misleading time and again in most surveys I read but let's go step by step. For the time being. Here's latest chart update. Anybody got anything to say you're most welcome to fire off anything you wish in the comments, as I've begun to reveal more of my initial premise.