A few days ago, Kapitano revealed that Google’s demographic algorithm had him pegged as an 18-24 year old male who was interested in martial arts, shopping and painting. He suggested that other people might like to find out what inaccurate ideas Google held about them, and thinking that it would have similarly chucklesome but vaguely flattering misapprehensions about me, I decided to follow the link he provided, and settled back ready for a good laugh. But the smile of happy anticipation rather died on my lips when it turned out that google doesn’t think I’m a vaguely hipsterish guy at least 15 years younger than I actually am. Oh dear me, no.
In fact, as this screen grab makes clear, Google thinks I’m male, 65+, and into the law and politics in a big way. In other words, I’m not just considerably older than my chronological age, but really uninteresting with it. I guess I should be grateful for small mercies, and that my interests aren’t listed as Chartered Accountancy, or Vinyl Flooring…
Anyway, when I’d got over the initial shock of being perceived as multiple decades older than I actually am (and that took quite some getting over, believe me) I noticed something odd about this list of interests, which is that it bears almost no relationship to the kind of websites I visit most often.
In fact, I have to ‘fess up to something here, which is that I initially read Kapitano’s post and followed the link several days ago, and there’s been a delay in me getting round to writing this post. (I’ve been having a bit of a crisis, mental health wise, over the last little while and thought I ought to wait until I’d remembered where I’d left at least a couple of my marbles before I set about blogging.) And when I first went to the site, Google still had me pegged as a sexagenarian, but one who was into astronomy, hip-hop and search engine optimisation, which at least made me seem like an interesting sexagenarian. The only real trouble was that list also bore little or no resemblance to my browsing history. It’s not that I don’t or haven’t expressed an interest in either set of topics Google have associated with my cookie, but rather that they account for a relatively small amount of the time I spend online, and things that I spend a lot more online time doing aren’t included at all.
I don’t know why that is. Perhaps Google is less interested in identifying the full range of my interests than in identifying those interests that advertisers are most interested in targeting (even if the lists seem a bit random to me, perhaps they don’t to advertisers). Perhaps Google think it’s more useful for advertisers to know what I’ve looked at most recently than what I’ve looked at most consistently. Perhaps my cookie is being regularly reset without my knowledge. Perhaps my habit – formed in those distant, pre-Google days when search results were different every time – of navigating to sites that interest me by remembering the URLs and typing them in directly makes it harder for Google to know what I get up to online than it would be if I searched for the same things repeatedly. (And perhaps my habit of copy-pasting URLs rather than clicking on them directly has a similar effect.) Perhaps too much of my online time relates to ‘sensitive topics’, and is thus excluded (if Google reckon all sites – even entertainment and politics ones – that mention LGBT issues are ‘sensitive’, then that’s perhaps as much as 25% of my browsing discarded right there). Perhaps it’s just that Google’s algorithm for quantifying and classifying people’s interests on the basis of their online habits isn’t very impressive – although this would come as a surprise, since the effectiveness of their algorithms is a large part of why Larry Page and Sergey Brin are wealthier than Croesus.
In any case, I think there are a few conclusions I can draw as a result of this. First, the nagging fear many of us have that “Google knows everything about me” would seem to be unfounded, at least in my case; Google know so little about me they think I’m three decades older than I am, and a monarchist to boot. Second – and assuming they are keeping track of all the sites I visit – I have to congratulate Google for avoiding a presumption of heterosexuality; given the number of times I’ve typed “[name of minor male celebrity] shirtless” into a Google search box and gone surfing off in pursuit of the result, they might very well have presumed I was a straight woman. Third, on the basis of my results here, I can see why advertisers are increasingly shifting their spending to Facebook, since the demographic profiling there is, presumably, more accurate than Google are able to manage.
And my final conclusion is that I still for the life of me can’t work out why they think I’m 65+. Ok, so that’s not so much of a conclusion as a wail of irritated confusion, but still, I can’t work it out.
Part (most?) of what’s going on here, of course, is that I’ve been mistaken for someone who’s aged 65+, and that’s annoyed me, and I’m having a bit of hissy fit about it. But I am quite surprised that Google’s demographic profiling should be as crude as it seems to be – why should an interest in ‘law and government’ or ‘politics’ (or, for that matter, ‘arts and entertainment’) be associated with pensioners? Aren’t some younger people (people my age, but even younger, too) interested in those things too? I’m sure, given a coarse enough analysis, it can appear that interests break down neatly along age lines – young people are into texting and The Only Way is Essex; people in their thirties and forties are into school league tables and Masterchef; people in their sixties are into politics and Countdown – but that misses a lot of variation, and a lot of detail.
It’s perhaps the lack of detail I find most surprising. Even TV companies are able to break down their demographics in more detail than this – identifying programmes that appeal to groups of people sorted not just on the basis of age and sex, but also occupation, income, and social background – and are able to offer advertisers opportunities to target their ads at these specific groups. Given that the great advantage of internet advertising over TV advertising is that it’s supposed to be possible to target it much more precisely, it surprises me that Google is using cruder analysis. If I was an advertiser, I think I’d be quite annoyed if I was paying Google to target my adverts at people aged 65+, and it turned out they were being put in front of 38-year-olds. In fact, I think I might be almost as annoyed as a 38-year-old who’s just found out Google thinks he’s a pensioner.
Anyway, if you’ll excuse me, I’m off to spend the next 72 hours streaming dubstep and surfing websites offering careers advice to teenagers in a desperate attempt to persuade Google that I’m at least the younger side of fifty. And if that doesn’t work I’m going to try repeatedly typing “Wagwan, my brother?” into Google’s search box. I mean, that’s got to be worth a shot, right?*
* – Or is it wa gwan? Given the meaning of the phrase (literally “What’s going on?”, although it’s used as a greeting, like the phrase “What’s up?”), and its derivation (Jamaican English), the amateur etymologist in me is pretty sure it should be wa gwan. But wait a minute – do you think publicly speculating about orthographic niceties in relation to street patois makes me seem old again? I bet it does. In fact, I should probably have typed it as “wigwam”, on the basis that’s what autocorrect would be likely to make of it. (As opposed to my spellchecker, which wanted to change it to “wan wag”, which sounds like the opening of a poem about a jilted footballer’s fiancé: “Oh, wan wag, with your bridal tresses in disarray…”).