Dear Jeremy, What is big data?
Oh dear. Here we go again. I struggle to answer this question (and not for the first time, I might add) only because I continue to cling to the hope that, in attempting to answer it, I might somehow begin to understand it. What follows are a few hesitant assertions that may or may not be true. I’d welcome corrections – but only if they clarify.
Ten years ago, if you’d sent me a postcard by way of the Royal Mail reporting on your holiday in Torremolinos, I might have read your postcard and even put it safely away in my sock drawer. But it would never have become a part of big data.
Today, if you post exactly the same information about your holiday in Torremolinos on Facebook, it immediately becomes a bit of (or a very few bytes of) big data.
Unlike the information carried by that postcard in my sock drawer, your holiday information is now available to just about anyone in the world with access to the internet: and that’s about three billion people. (Or it was at eight o’clock this morning; it’s now several million more than that.)
So you shouldn’t be surprised if, very shortly after posting it, you’re personally invited to buy a certain suntan lotion, fly a budget airline to Málaga and browse through an online catalogue of sunglasses. Information gleaned from big data will have been put to commercial use by three digitally sprightly companies.
But nor should you be surprised if none of the above happens.
The big thing about big data is this. It’s made up of zillions of very small pieces of data. Facebook alone generates 500 new terabytes of data every 24 hours. (And one terabyte, as I’m sure you already know, is a thousand billion bytes – or thereabouts.) So the chances of an airline stumbling on your Torremolinos itinerary are small; except, of course, that increasingly intelligent automated scanners can sieve through a few terabytes of data in seconds in the search for any clue that might be of interest to airlines, shades or suncreams.
Companies are very excited about the potential value of big data. They don’t have to wait six months for the research report to find out what happened; they can learn what’s happening while it still is.
The competition between companies is not to produce still more data but rather to learn how to use what’s already there. And it’s mounting by the millisecond.
According to the International Data Corporation, less than 1 per cent of the world’s unprotected data is actually scanned and analysed. But even 1 per cent is huge. Big, in fact.
Dear Jeremy, I’ve got a smart new chief marketing officer running my marketing department and I’m now thinking of bringing all our advertising in-house. We’re a pet-food brand with a decent ad budget and I reckon, with a few more smart hires, we could do the whole creative thing ourselves. Do you have any advice?
You might make it work for a year or two – but why?
Why try to buck history? Lintas opened up in 1899 as the in-house agency for Lever Brothers and gave it as good a try as anyone. There are at least two good reasons (and one bad one) why the in-house agency will never make long-term sense for a company such as yours.
Good agencies become good and get better through working on the widest possible variety of different businesses; that’s how they learn about advertising rather than about the pet-food market (that’s your job).
You’ll never keep the best creative people (or the best planners or suits) if all you can feed them is pet food. So next thing you know, you open your doors to other clients – and spend most of your time fighting off the reputation of being the pet-food agency. So you’ve now got the worst of both worlds.
The best clients want a bit of honest truth from their agencies. It can be hard enough to risk unpopularity with a client when you’ve got 20 others. An in-house agency has only one. How can you expect detachment from people when you’re their only paymaster?
And, after five poor sales periods, when your keenest competitor has put on two share points and their agency has pranced back from Cannes with a category Lion: who are you going to fire?
Jeremy Bullmore welcomes questions via email@example.com or Campaign, Teddington Studios, Broom Road, Teddington, TW11 9BE