No, Big Data firm, the UK isn't teeming with UBER-FRISKY GIGOLOS

Prostitution + official GDP figures = buckets of FAIL

This is self-evident hokum. Here's why

Maybe those numbers aren't quite right. But think how different they've got to be if we're going to have the 30 million odd men (or 28.5 million hetero men) of this country keeping 60,000 working girls gainfully employed full time and the other 1.5 million keeping 45,000 blokes in tea and crumpets. It's just not quite going to work, is it?

Whatever would June Whitfield say if she was served a cup of tea from a hyperactive male prostitute? Not that we're suggesting this is happening in this Orange press pic, obviously

But that is what import.io has assumed: that those 45,000 male escorts are all having 25 encounters a week at £50 a pop, matching what the ladies are doing. Quite apart from anything else performance issues would come into play: Viagra's pretty good stuff but not, perhaps, that good.

My point here, though, is not really to discuss the ins and outs of the sex industry: rather it's to point to something that economists struggle with and which this big data stuff might have to consider too. Data on its own isn't enough. You also need theory and then you need to have a reality check. Does this finding that we've reached by using our data actually accord with what we know about the subject under discussion? Is it likely to be true? Is it even believable?

Run our numbers of male prostitutes against the numbers for the gay population, use that 25 punters a week, 100 a month, tricks turned and what do we reach? Well, actually, 45,000 people turning 100 tricks a month means, assuming each punter only rocks up once a month, that there's a population of 4.5 million as customers. And yet the gay and bisexual population of the country is put at some 1.5 million men.

We can only believe import.io's numbers if we're prepared to believe one of several different things. The number of women buying sex is very much larger than we generally think it is or nearly all gay men are buyers of sexual services almost weekly – neither of which really seems very likely: and that's the reality check that should have been done before they released these numbers.

Quite where the truth lies I have no idea – my suspicion would be that many of those men offering those professional services are very much part-timers – but my point about the perils of big data, I hope, still stands. However delightful it is to crunch the numbers, your results do still have to accord with both logic and observable reality. These don't – so there's something wrong with the number crunching. ®