Predicting the future — once the realm of election analysts, sports bettors, and charlatans — is now a central part of digital marketers’ resumes.
To help marketers stay ahead, personalization technology provider Sailthru is today joining the ranks of future-tellers with its announcement of the addition of the Sightlines solution to its core product.
The new service, now in private beta with a general release planned for early next year, is designed to predict a range of customer behavior — what the customer will buy, how much she will spend over a time period, what content she will like, even whether she will open certain kinds of marketing emails. All of the customers are known to the brand and are individually identifiable.
“The real difference [from the previous Sailthru product] is we’re now predicting the future,” CEO and cofounder Neil Capel told VentureBeat.
Sightlines is intended to ingest “every communication touchpoint” with the customer, the company’s chief data scientist Jeremy Stanley told us, including interaction with the brand’s Web and mobile websites, apps, email messages, and social media. Although user behavior on mobile devices can be different than on other devices, this first iteration of Sightlines does not break out mobile predictions. But Capel said that was in the works.
“There are three key ingredients to make this work,” Stanley said — comprehensive data, predictive analytics, and “an ability to act on the predictions” through messages directed at the individual.
There are a variety of predictive analytics tools that integrate with marketing platforms, such as Infer, and some marketing suites that tout their onboard predictive marketing capabilities, including Adobe Marketing Cloud or Salesforce‘s ExactTarget Marketing Cloud.
But the New York City-based Sailthru contends that it offers the only platform where all three are integrated, easy to use, and available in real time. Stanley also said that the predictive capabilities in a platform like ExactTarget are really recommendations.
The difference, he said, is that recommendations make suggestions about other content or products, based on content or products the customer has engaged or purchased.
“Predicting,” Stanley said, means “this customer has this [percent] chance of doing [something] in the future,” including “opt out of communications [or opening] an in-app message.” In other words, a recommendation compares like products, while predicting forecasts a range of future actions.
Getting the future right can mean a big difference to bottom lines. Stanley told us that Sightlines can pinpoint “the top ten percent of predicted [known] users who will generate 90 percent of the client’s revenue in the next 30 days,” from the brand’s pool of identified individuals.
Sailthru said its three million predictive models are rebuilt daily from the newest data, so if an individual keeps defying the prediction, the prediction will be modified.
If all this feels quite deterministic — you don’t know what you’re going to buy, so how does the brand? — get ready for more: Stanley said he expects predictive accuracy to grow, especially as the Internet of Things offers a zillion more data points about what you have been doing.