Display advertising has always had trouble proving its business value due to the fundamental flaw of focusing almost exclusively on its "last-event" role, so how did Waitrose go about tackling this? Infectious Media's Attila Jakab explains.
Four in five marketers we surveyed recently, admitted they still have problems measuring the effect of display advertising. I believe this is because most advertisers and their agencies are still using "last-event" measurement - be it last-click or last-view - as the benchmark for how display ads are deemed to have performed.
This methodology is flawed simply because it ignores the entire customer journey online - crucially the brand awareness role display ads play, just like TV, print or radio ads do. Why should display be saddled as a last-event tool just because that element can be measured better than any other ad medium?
What's potentially more damaging to the bottom line is that, by mistakenly focusing on how the last-click performs, it encourages advertisers to over-invest in activity at the bottom of the sales funnel rather than ignoring the top and wider parts of the funnel that can deliver much better return on investment.
So, we teamed up with Waitrose, an innovator when it comes to digital marketing, to tackle the issue and to more accurately define the real value of display ads.
We agreed that the true value of an ad could be defined as sales that wouldn't have happened without the customer seeing the ad. The really challenging part of this is the absence of an "attribution" model that measures the various touch-points - both online and offline - a person encounters between being made aware of a product and purchasing it.
Working with Waitrose, we came up with a better way to measure the value of display using one of the industry's hottest and most divisive topics - ad viewability - and combining it with other data points.
The IAB defines an ad as being viewable if 50% of it is in view for at least 1 second. It seems redundant to say it, but a fundamental element of changing customer behaviour is that a person actually sees the ad - you probably wouldn't believe the number of advertisers that don't track this simple measure.
The test measured the behavioural differences - such as website behaviour and purchasing - between people who were shown at least one viewable ad and those shown at least one non-viewable ad in the same campaign. A non-viewable ad, for example, could be one lower down the page which a user never scrolled down as far.
Our initial finding was surprising and initially discouraging: sales from customers who were shown viewable and non-viewable ads were exactly the same. In other words, a customer would have made a purchase regardless of whether they'd seen the ad or not.
Sales from customers shown an ad has been the industry standard for measuring the success of display for a number of years. This discovery exposed this standard as fundamentally flawed.
However, after extensive testing and optimising - something that could only be done programmatically - and analysis, a breakthrough was made.
The factor that had the biggest impact on sales was the number of viewable ads a person was served.
To be more specific, an uplift in sales occurred once someone was served with five or more viewable ads in one month. It's worth also noting that the uplift plateaued after being served 10 viewable ads in a month - in other words, there was no incremental benefit in serving more than 10.
This "5+" ads finding revealed that serving viewable ads to people at a low frequency level was just as wasteful as serving non-viewable ads to users. To put it another way, it's a complete waste of money to serve four or less viewable ads to a person in a month. It's only worth it if you're going to hit them at least five times.
Once this discovery was made, it meant the campaign optimisation strategy could be changed to focus on delivering the ideal number of viewable ads to a potential customer. The result was an instant increase in the number of incremental sales as the campaign was optimised.
This innovative approach enabled us to measure, report on and optimise directly towards sales that could be directly attributable to advertising. By disregarding users on whom the advertising had no impact, Waitrose ensured sales were maximised and wastage caused by non-viewable and low frequency ads was minimised.
As promising as these results are, there are still more questions that need to be answered before the true value of display advertising can be measured. However, it shows that moving away from solely measuring the "last-event" is a very good first step.