Said Mr. Jones: "It's incredible how ever-impressed people get by their own social-media statistics," noting that one company's 500 likes could be another's 5 million likes. They were down on the lack of context, too, provided for some of the metrics. For example, saying that website traffic jumped 20%, but not providing a time frame of when that happened, and how it was directly linked to the advertising campaign.
Attribution in advertising is going through a bit of an odd phase at the moment. There are a few reasons for this. Clients are asking for more proof that their money is being well spent. New media are introducing stacks of new metrics that need to be learned and tested. And the current interest in social and integrated campaigns mean there is a lot more to measure. Suddenly, measurement and attribution is the thing to be talking about.
One side-effect of this is that a lot of people who like to say smart things about advertising are trying to out-do each other in the conversation around metrics that matter. Social channels have become a gold-mine for these sorts of arguments. You'll have heard the kind of thing. Is sharing important? Do 'likes' matter? Should you build fans? Does online video get results? And so 0n, all of them with many contradictory answers.
Obviously, the only answer that matters is: 'It depends.'
Do fans matter? Well, it depends what you're trying to achieve. If you're aiming to convert people to a sale as fast as possible, maybe not (though maybe). If you're trying to build a regular audience for communication over the medium term, then maybe (though maybe not). Does online video get results? Well, it depends on the objective. Like all advertising - actually, like all spending. Whether something is worth paying for depends on what you want it for and what you'll get out of having it.
So far, so simple - but it's amazing how many conversations about this sort of thing seem to be motivated by people wanting to have a flat-packed smart answer, rather than saying, each time: 'I don't know yet. Let's take a step back. It depends.'
But back to attribution. In a world where marketers say 'it depends' more readily (good thing), they also need to say what it depends on more often. When your 'it depends' becomes an 'it's impossible to say', you're in trouble. Then you get desperate, and start attributing any positive result to your activity. Then the Cannes judges will shout at you.
So do your clients a favour and start testing hypotheses. Start keeping records - baselines of what people are saying about your brands or products; traffic to your sites; how fast your fans grow; sales; things you don't know if you'll need or not. Learn to watch for patterns. Learn about correlations and how to measure significant change in trends. Understand variables, build ranges of estimates, plan for different outcomes. Count things. Tell stories with numbers.
That way, when someone asks you how well a campaign is expected to perform, you can avoid either an awkward silence or an impossible prediction. 'It depends' won't win you any blogging awards, but it'll mean you're giving better advice. Which is what you're here for, after all.
# Alex Steer (21/06/2012)