We humans are an incredible bunch. Our ability to quickly assess situations and assimilate data to come up with solutions to problems is amazing. We apply that ingenuity daily, whether it’s calculating in a split second all the complex physics and geometry it takes to catch a ball, safely navigating a busy motorway in the rain during rush hour, or firing a reusable rocket into space. And landing it again.

This same versatility has led us to create computers and new forms of communication which have driven exponential growth in the information we share and the data we create on a daily basis. In turn, we’ve invented new and more powerful computers and techniques to help us grasp the potential of that data and apply it to better solve even more problems.

After finance, the $500bn global advertising industry is possibly the fastest-growing source of data on the planet. Insights into consumer preferences and behaviours are driving billions upon billions of transactions between brands and publishers daily as advertisers seek to reach and influence a growing and more prosperous online audience. In recent years, an entire industry of intermediaries has sprung up to fill in the skills gaps and seek their share of the money flowing through the system.

However, the system is far from perfect. The way most companies measure online advertising campaigns today - with all their sophisticated complexity and nuance - still largely relies on measures developed for the early, simpler days of search advertising and the rise of Google. Back then, it was pretty easy to see whether an ad was relevant to what a consumer was searching for: they clicked on it.

Using a crude measure like clicks and its close relative ‘last touch’ to measure modern ad campaigns - and plan future investment on the back of those measurements - is a bit like betting your life savings on a company’s stock because it went up one day. It’s a positive indicator, but there are lots of other signals you should be considering before placing that bet. Yet, many companies are investing precious marketing budgets based on one signal out of the dozens or hundreds that they could and should be looking at.

This mismatch between activity and measurement is creating poor data which, for the lack of anything else, marketers are forced to act on. The age-old rule of garbage in, garbage out still applies. Bad data is leading to bad decisions and advertisers are becoming increasingly disillusioned by what they’re seeing. Naturally, they’re pushing back against what look like murky practices, and they might be right. In the dark shadows, fraud thrives.

As a result, trust is being compromised. If the power and potential of new innovations like artificial intelligence (AI) risks are to be unrealised, then something has to change.

In order for the advertising industry to avoid being bypassed by brands seeking more transparent and direct engagement with their audiences, or increased legislation from governments, these issues need to be fixed. By doing so, we will free up a huge amount of human energy and imagination currently focused on tackling issues like ad fraud or brand safety. Instead, we can direct it to doing what most of us got into this industry for in the first place: using human insight to find more effective and creative ways to connect brands with people.

Understanding the consumer is at the heart of advertising. As humans, we tend to think about data in dimensions we’re used to, such as age, gender, interests, and so on. We can quite comfortably deal with a certain number of variables for a certain number of people (though, for some reason, buying Christmas presents is still difficult). With more data we can build more accurate pictures of consumer behaviour and preferences. With the data we now have and the speed at which it’s growing, we need machines to help us effectively do this.

Computers aren’t limited in the same way as we are. They’re able to instantly assess millions of potential variables and do it for millions of consumers, constantly updating their models and predictions based on new inputs from each ad campaign that consumers might be engaging with. This is what we’ve built at Quantcast over the past 11 years: an AI-driven live model for the internet-based on real human behaviour that advertisers can base decisions on and that can be calibrated in endless ways, operating at huge scale and speed.

Today, humans are still playing a large role in working out which sites to advertise on, with what messages and with which combination of ad space, bought at what price, when, and at what frequency for a particular campaign. With the right inputs, computers can take on all that tedious, repetitive, and precise work and do it better. This allows people to focus on the human side of advertising: developing compelling new creative, innovative media strategies and so on that require much more high-level thinking. The technology already exists in order for us to do this. It just requires an update in our thinking about measurement in online advertising for us to get out of the way of our own progress.

The result for brands - and for the industry as a whole - will be more effective campaigns allocating client budgets where they drive the most business impact, not just the most clicks. From a situation today where everyone is competing fiercely for a 10 percent gain in efficiency, we could see regular leaps in campaign effectiveness ten times greater than before, and all with a much more human touch.


By Ben Murphy, UK managing director at Quantcast

GDPR Summit Series is a global series of GDPR events which will help marketers to prepare to meet the requirements of the GDPR ahead of May 2018 and beyond. Further information and conference details are available at http://www.gdprsummit.london/

comments powered by Disqus