Dropping Stereotypes: How We Go Beyond Targeting

By Sharon Malaby

Relevant defined in the dictionary means, “bearing upon or connected with the matter in-hand”.  In the world of advertising, relevance means finding out what matters most to people in order to deliver brand messages that are designed to be interesting, engaging and ultimately behavior changing. For many advertisers this is what they are striving for – a true connection between the brand and the person.

Traditionally, the method for pinpointing relevant ads has been done through targeting. What I’m referring to is using predefined explicit targeting such as demographic data, i.e. your age/gender, your location or your device.

The challenge with this method is that a lot of assumptions are inevitably made through the use of targeting. We all get boxed in and stereotyped by our age and gender or where we’re from.

Targeting assumes that because you belong to this group of people – you will then share the same experiences and interests as each other. This simply is not the case in the world we live in today.

All mums are not the same. All men aged 18-35 are not the same. Invariably, they all have different interests, too. Targeting and audience segmentation looks at just one aspect of who you are. Or rather, the generalisation of who you should be based on just a few hard facts.

At the same time, the media landscape is becoming exponentially noisier. It is tempting for advertisers to get into a battle of who can scream the loudest in an effort to get people’s attention. But a meaningful and long term bond cannot be built this way. Instead, what if marketers could actually predict when someone would be most receptive to their brand or product based on a multitude of behaviours, and therefore deliver messages at precisely the right moment?

This is exactly what drove our Research & Development (R&D) team at Widespace to build: predictive technology for mobile brand advertising.

We have always believed that relevance is the foundation to successful advertising; an ad that is relevant suddenly becomes informational and not just noise.

If you are addressing the right people at the right time with the right message, you save a lot of effort trying to convince someone that is not willing to listen. That same individual may be just right a different day in another context.

You may be asking: how can we possibly know what people are interested in and what is in fact relevant for them? People’s interests and attitudes change rapidly, sometimes over one day, and depend on many factors. Traditional targeting methods cannot address the dynamic and fast-changing nature of human interests. In order to predict interests we must first be able to accurately measure when someone likes an ad. I’ll give a quick description of how this is done so bear with me….Clicks are the most obvious indication of interest – which is what most delivery platforms are based on today. However our research shows that only 4% of people actually click on ads. Our research surveys also indicate that 9 out of 10 people who express interest in ads still do not click on them. Hence other data sources must be used as indications of interest.

We therefore must abandon the click in favour of better and more accurate interest indicators. Widespace technology uses a combination of swipe patterns on ads and content, relative view time, content history and interactions, such as shaking or talking to an ad, all to reveal a person’s ad interests. In essence we are using Widespace ad formats as little data collecting machines. Our data team now has an universal way to measure ad interest because 100% of people view ads and swipe on mobile.

Screen Shot 2016-01-12 at 11.16.25
Click to Play Video

Using all of this data, our R&D team then set out to build a model which predicts ad interests. For instance, by looking at your behaviours with previous ads we can predict precisely when you will be more receptive to a fitness message. The complexity of this is hard to explain so we have a video to help: Beyond Targeting. Early findings show that we are headed in the right direction – which is extremely exciting and promising. Our A/B tests have already shown an uplift of 13% in ad view times and 36% in interactions on a campaign level when compared to the old methodology of targeting. Ad closure rate has also reduced as a byproduct of better delivery, even though we are not actively optimising on close rates.

Creating the perfect prediction is of course an ongoing challenge. The prediction power of our technology is constantly being improved as we become smarter.

We are continuing to refine predictions in an effort to create the perfect match between people and ads.

Targeting has been the old tried-and-true way of finding the audience for an advertiser’s message. But to be even better at delivering effective advertising, we knew something more had to be added. By serving the right ads, less ads are needed and ad fatigue can be avoided. We reduce ad wastage while boosting ad relevance and ultimately improve metrics like recall and purchase intent. Hence creating a better world of advertising where advertiser, consumer and content provider all win. A positive and valuable exchange between brands, people and content owners is built as a result. Everybody wins.