Those of us who have been watching the surge in mobile devices in everyday life knew that the rise of real-time bidding (RTB) in mobile advertising was simply a matter of time. What is less obvious, but just as inevitable, is the evolution of the targeting tactics that drive mobile RTB. Advanced predictive modeling technologies significantly out-perform traditional behavioral targeting methods, unleashing the maximum value RTB for mobile.
RTB itself works because it offers immediate value to advertisers and consumers. Mobile RTB allows the advertiser to cherry-pick target impressions out of a pool of hundreds of thousands of impressions per second. With this basic understanding of mobile RTB, the standard way of buying media and serving ads across a specific site's entire inventory just doesn't make a whole lot of sense anymore.
In most cases, the targeting that powers mobile RTB is based on a user's online behavior, or behavioral targeting (aka audience targeting). An advertiser might want to serve ads to someone who has visited the specific sites or content before, or to someone who has shown specific behaviors in their browsing that fit an audience profile predetermined as the target for the campaign.
Though behavioral targeting may have been considered effective enough in the past, there are inherent limitations that are too fundamental to ignore any longer:
- Targeting individuals who have done something that reveals their interests is not the whole story.
- This data used to target your customer is narrow and often fragmented, which offers limited real advertising value.
- Little or no correlation is made between the behavioral data collected and other conditions (i.e. location, time of day, weather, etc.).
The amount of wasted impressions and missed opportunities using current targeting methods is astounding. Why would you ignore millions of mobile users who haven't yet visited your site or displayed a specific behavior? Many of them may actually be prime for your campaign. Predictive targeting offers a way to find them, and you can't afford to ignore it.
Alternative approaches based on innovative technologies can now "predict" which user is most likely to respond to an ad based on a wide range of targeting parameters instead of just one set of specific behavioral data. Predictive targeting uses sophisticated data mining techniques to calculate the probability that an impression will result in a conversion. The process starts by gathering first party data received from the initial ad request, then enriches this data with third party data, historical conversions, ambient data, and ultimately matches the fully enriched impression to a brand and creative -- all in real-time.
Mobile RTB platforms enable advertisers to look at the shared attributes of those who have converted, and the varying factors that resulted in the conversion. Data such as socio-economic factors, weather, and much more reveal a pattern that allows us to predict other people who match these situational data points, and will be far more likely to convert. We can then target impressions through the real-time bidding exchanges, and reach a maximum ROI for advertisers.
Why is this new way of targeting using big data such a game changer? Predictive Targeting allows you to find new customers, users who would most likely otherwise remain off your radar. This isn't to say that behavioral targeting can't find new customers. However, it's inefficient at best. You can throw 1,000 (behaviorally targeted) darts and hit the bulls-eye maybe five times, or laser focus your (predictively targeted) darts and hit the bulls-eye 15 times or more.
Adding to the power of predictive targeting: it continuously learns. After initially hitting the target 15 times, your campaign will increase in accuracy and precision through out the rest of the campaign. My prediction? RTB driven by predictive targeting will continue to gain traction and become the preferred method of reaching a brand's audience in mobile.