It was only a few years ago when the ad industry was being confidently promised that digital marketing and big data were going to change everything about how we measured the effectiveness of our efforts.
In 2008, the first CMO Survey in the US found that the biggest planned growth area in marketing budgets was in "marketing knowledge", including research and effectiveness measurement, which was set to grow at five times the rate of marketing spending overall.
Fast-forward to the latest survey in 2016 – we see marketers spending 6% of their budgets on data analytics and anticipate that they will have to double this over the next three years. Despite this, only a third of marketers say they can prove the business impact of their marketing spend.
We’ve moved from a world where marketers had to make judgments with too little data to one where we have to make decisions with too much. For every one metric we had to work with, we now have dozens or even hundreds, letting us track every impression, click, conversion, "like", share, retweet… you get the idea. An ad-served world is a world of numbers of varying degrees of enlightenment.
This oversupply of data means that, for almost any campaign, it’s possible to find some metric that went up. Didn’t drive clicks? Take a look at "likes". "Likes" underperforming? See what happened to shares…
It’s easy to become the analytical equivalent of a conspiracy theorist, forever seeing patterns that aren’t there to justify a narrative that you would have come up with anyway.
For agencies, too, a naïve approach to data-driven marketing can lead to the nerve-shredding sensation of having to be judged on 15 performance indicators at once (all of them, inevitably, "key").
I don’t advocate that we switch off the data hose and go with our intuition. New kinds of data give us vastly more insight into how people consume and respond to advertising in detail. But we urgently need to move beyond just hoarding short-term metrics. It’s time to rehabilitate that old notion of "marketing knowledge".
To put it another way, we must move from metrics to models – from drowning in isolated data points to using analytics smartly to understand the underlying patterns and norms in advertising performance, quantify the business impact over the short and long term, and give us a more reliable guide to the future.
At Maxus, we’re taking some of the pillars of marketing effectiveness – KPI-setting, benchmarking, segmentation, econometric modelling – and fusing them with new, fast-moving data sources and methods such as cluster analysis and digital attribution. This helps us take advantage of what an ad-served world can offer without getting lost in the noise.
If we are to give ourselves the freedom to innovate with confidence, advertisers and agencies also need to look beyond our own data. That means combining our evidence with knowledge created elsewhere. We don’t have to prove everything from first principles on our own brands. While methods such as experimental testing are useful, they do not replace the experience of others.
Initiatives such as Newsworks’ research programme are hugely valuable because they give us clearer insight into the general principles of how advertising works – in this case, a view of the role news brands play in the media mix to generate short-term sales and long-term brand health.
In 1921, CP Scott, editor of The Manchester Guardian, wrote: "Comment is free, but facts are sacred." A century later, we find the tables turned: data points, the raw material of facts, are available almost without cost; but analysis, the root of informed commentary, is an investment worth making.
Alex Steer is the head of technology, effectiveness and data at Maxus UK. He speaks at on 12 October.