The humble mathematical algorithm has become a mighty tool in programmatic buying, finding itself at the heart of online advertising and the search to optimise value from publishers’ inventories. And it is certainly important as a means to create order out of the unfathomable amounts of data online. However, people are putting too much faith in what they expect an algorithm to do.
The issue is that with maths there is only one right answer – and digital media doesn’t work like that. It’s not just about placing an ad on the page, it’s about people. As David Ogilvy said many years ago, the function of advertising is to sell. i.e. Getting an ad in front of someone five or six times is not enough; the ads need to ultimately create a sale and to do this they need to be relevant with the people you are talking to. They should either desire that product/brand right now, or maybe aspiring to buy it in the future.
To create this relevancy, algorithms are simply not enough on their own. The essential ingredient required if you want to start achieving response rates that are above the sickly standards that seem to be the norm is – insight. Without the input of human intelligence, algorithms are little better than billboards; you hope that if enough people see the ad then something will stick. However, funnel data down into an analytic engine and then add some brain power and you can start to learn creatively and intelligently about what factors make a real difference to response. After all, you need to be human to understand human nature.
Also, however good your algorithm is, one size will not fit all in Real Time Bidding (RTB). Every brand and campaign is different and human input and instinct can make the subtle changes that will create a big impact on campaigns. There’s more to advertising than ever before – a myriad of brands, different screens, on and offline data, transactional data, credit data, unique users, repeat purchasers. All these elements need to be understood as they have an impact on driving acquisition and the brand. Simply putting them into a machine and expecting the right answer to come out the other end is madness.
Especially when the algorithms being used are often based on US data. We may have a special relationship with America, but we think and behave in a different enough way to our US friends for this to be a significant issue in digital advertising. There are probably enough differences between North Londoners and South Londoners to impact on campaigns – let alone across different continents!
It’s time to stop following the American RTB trail blindly and create our own programmatic path in Great Britain. The UK has started to change the way it sees things and that is having an effect on how our cousins look at product development. We do need algorithms (although these need to be based on relevant UK data) but we need these as our starting point. We then need to add on our own human interpretation to what the numbers are telling us if we want to raise the bar in terms of response and conversion.
The technology is already available to drive cars without human input. But do we really think that will make the roads safer?
By Donald Hamilton, Managing Director of Zodiak Advertising.
PrivSec Conferences will bring together leading speakers and experts from privacy and security to deliver compelling content via solo presentations, panel discussions, debates, roundtables and workshops.
For more information on upcoming events, visit the website.
comments powered by Disqus