Overcoming Retargeting Challenges Through Predictive Segmentation

In preparation for the holidays, I recently took a trip to a local electronics store to buy a gift for my son. While I have visited the store frequently in the past to browse their selection, I had yet to make a purchase. As I walked through the entrance, I was approached by an employee who was passing out coupons for a discount off their hottest item -- coincidentally, the item I came to purchase.

As I considered this offline remarketing tactic, it got me thinking about its online equivalent, which we all know as retargeting. Retargeting platforms serve specific messages to consumers based on their previous actions, such as visits to a website, that did not result in a sale or conversion.

While retargeting has emerged as a popular form of advertising, it isn't without flaws. First, the retargeting universe is relatively small, since it includes only those people who have been to your site but have yet to convert. But the bigger issue is that retargeting influences the people who were already interested or influenced by the marketer doing the retargeting, just as I had already made my decision to purchase the item for my son before getting the coupon.



Other issues arise when marketers use last-click methods to measure the effectiveness of their retargeting efforts. Using last click, retargeting ads receive full credit when they are the last ad presented to a converting user, regardless of any other touchpoints that may have contributed to the eventual conversion. Marketers who serve retargeting content to users who previously visited their site often struggle to understand the incremental performance of those ads in driving conversions, versus what part of those conversions would have occurred regardless.

So how can marketers better reach likely converters? Predictive segmentation is one viable option for reaching those consumers that have a higher probability to convert. Put simply, predictive segmentation enables marketers to identify and score people who are more likely to convert based on previous behavior, and deliver relevant, targeted ads to those high-scoring people.

Here's how it works: marketers using any sophisticated measurement system is able to track the behavior of their entire universe of users based on propensity to convert. These advanced modeling techniques are able to correlate the behavior of certain users with future conversion activity. Because marketers understand the value of each user and their inclination to purchase, they can use that information to create a group of “likely converters” and deliver the most effective ads to them in order to entice conversions and achieve maximum lift.

While predictive segmentation sounds similar to retargeting, the two are actually quite different. Retargeting only targets a small percentage of people who have recently demonstrated interest in a product or service, by visiting a brand's website for example. With predictive segmentation, the potential pool is much larger, since segments are based on the probability that a given user will convert in the future. Moreover, using predictive segmentation, marketers can select as large or small a sampling of individuals to target based on their available marketing spend. For instance, marketers with smaller budgets can target the top percent of those “most likely to convert” for the most effective results. Those with larger budgets can target as high a percentage as they'd like – broadening their reach of potential conversions well beyond retargeting methods.

Using predictive segmentation, marketers don’t have to wait until I enter the front door to offer a coupon. They can offer relevant, appropriate offers to me because I have exhibited shopping behaviors that demonstrate my likelihood of converting  -- like slowing down as I pass their store window  -- thus incentivizing people who wouldn’t have walked through their door to take that next action.

Next story loading loading..