Alex Steer

Advertising effectiveness, analytics and strategy / about / archive

Are we underselling big data?

1257 words

As we get into the New Year, I seem to have read about a hundred blog posts that all say the same thing. There's lots of hype about big data - but just having data isn't enough; it's what you do with it that counts.

To which my reply is: yes, well, obviously. That would also be true of any marketing asset, from a website to a tracking research study to some media space. Merely noting that people are excited about something, and some people will use it badly, is a lazy way to try to sound smart.

So I'm going to sound perverse instead. I think, despite all these blog posts, that big data is being undersold to marketers. Here's why.

There's a gap between the hype and the sales pitch

There's a difference between hype and sales. Hype is a precursor to sales, but sales are the closer. Having lots of people excited about something (be it big data, subprime mortgages or Dutch tulips) is a precursor to making a killing when you go round and persuade them to pay for it.

But whereas sellers of snake-oil or South Sea shares tend to be more excited than the wary customers to whom they're flogging their wares, the people selling big data services - getting the SOWs signed, agreeing the terms of contract, receiving the purchase orders - seem curiously less enthused by its potential than their would-be clients.

There's an odd gap, in other words, between the hype surrounding big data, and the things big data providers are selling.

The dismal quadrant

There are a lot of technology companies springing up to serve the marketing industry, as this diagram from the Chief Marketing Technologist blog shows. But broadly, I find you can cluster them into four big blocks, based on the answers to two questions:

  1. Do they sell raw data, or services informed by data?
  2. Do they promise to improve the efficiency or effectiveness of marketing activity?

Based on that, the landscape breaks down something like this:

data-landscape

Three of these quadrants are pretty straightforward.

  1. Greater efficiency for raw data. The world of data management platforms and cloud-based storage.
  2. Greater efficiency for digital services. The world of digital delivery and asset management platforms.
  3. Greater effectiveness from raw data. The world of syndicated research and heavy-duty analytics.

All three are sold quite well by the people who sell them. But when marketers get excited about big data, they're not talking about any of these three. They're talking about the promise of the fourth quadrant: Greater effectiveness from digital services. To put it another way: Can data-driven services make my digital marketing better?

And by 'better' they don't mean more efficient. They mean more interesting, more powerful, more pulse-racingly good. The kind of 'better' that wins awards and makes clients famous and changes people's minds for the long term. The kind of 'better' that makes good people want to work in this industry in the first place.

That's what they're asking us for. And our answer?

Targeting.

Thomas Carlyle called economics 'the dismal science'. Targeting may be its successor - accurate, useful, but not exciting. Clients want services that realise the hype around big data - full of surprising truths, powerful insights, previously unseen revelations and ways of working that are completely unlike the often frustrating reality of trying to look after the brands by which large businesses grow and thrive. And we give them the ability to swap one banner ad for another based on some fairly simple rules.

Putting the 'big' back

We're selling big data with small ambitions, and that needs to change. In large part, it happens because the people rushing into the sector aren't marketers. To steal from Tom Morton's outstanding recent presentation, many of them 'haven't a brand strategy or creative bone in their body'.

This is the territory that agencies should be occupying, and for the most part they're stubbornly refusing to. When was the last time you heard an agency describe its main aim as making digital communications with more impact, instead of some dismal old rubbish about conversations? Instead of competing on unachievable claims about creativity - which, let's face it, should be table stakes for a creative agency - agencies should be running headlong towards this empty, underserved but incredibly important quadrant. Instead, it's populated with companies that think that marketing effectiveness boils down to nothing more than optimisation after the fact.

It's time to stop acting as if every call for greater focus on effectiveness is an attempt to hammer nails into the coffin of creativity. It isn't, because the two go hand-in-hand and you can't be serious about one without the other. But clients need to be demanding to work with partners that are equally serious about both: not agencies that go 'la-la-la' every time questions of analytics and scalability are mentioned; and not tweakers-at-the-margins who see the content of marketing creations as largely irrelevant.

And it's time to reassert the distinction between strategy and tactics. As I review the kinds of supposedly strategic advice a lot of clients are getting from the technology sector, I see little more than a set of tactics backed up by case studies of their previous effectiveness, with little reference to the context in which those effects were generated. That's the model-based way of selling technology consulting services at its worst: use technology, ???, profit. The best strategic organisations - be they consultancies, ad agencies or tech companies - know that strategy is based on a need and a theory of change, and leads to a set of actions that make a business's marketing output make more sense - sense for the business, and sense to its would-be customers.

So in short, yes, I think we are underselling big data by using that term to describe some very small, tactical solutions to problems that we're not defining well enough. Clients should be asking partners to use big data to solve big, specific, thorny problems, and not be satisfied with any solution that isn't both empirically grounded and rooted in a proper understanding of behaviour.

It may be that we need to rip up some organisational models to do this - the distinction between ad agency and tech company, for example. That won't be easy, but it will be worth it, and it's what clients should be asking for this year.

# Alex Steer (28/01/2013)