Oracle’s Responsys Adds Machine Learning-Driven Personalization Via Boomtrain Integration
Email marketers on the Oracle Marketing Cloud can now skip creating multiple email variations and segments when they want to target individuals.
Personalized notification platform Boomtrain announced today a new integration with Oracle’s Marketing Cloud that could make marketers’ lives a bit easier.
The integration provides a predictive personalization platform for email marketing through the Oracle Cloud’s B2C Responsys component, which enables marketers to conduct campaigns via emails, website ads, text messages or social media ads.
Boomtrain specializes in machine learning-based personalization for email and push notifications, and this is the first machine learning solution for Responsys. Machine learning uses pattern recognition to learn from previous examples, thus improving effectiveness as it acquires more data — without being specifically programmed for the improvements.
“This is the ‘easy button’ for marketers,” Boomtrain VP for business development Neej Gore told me.
Without Boomtrain, he said, a marketer looking to send out an email newsletter would first prepare a dozen or more variations on the Oracle platform, create audience segments that grouped recipients according to related profiles or other data, and then send out the right variation to the right segment.
With Boomtrain, there’s one email that has places for different kinds of existing content — product info, white papers, links to articles, photos and the like. Here’s a screenshot of Boomtrain used inside Responsys:
Boomtrain focuses mostly on publishing and ecommerce client companies, so this kind of selectable content is not uncommon in the emails sent in those industries. A publication might send out recent stories, for instance, or a retailer might send products on sale.
Gore said the marketer drops some Boomtrain code snippets into the newsletter, which then grabs the appropriate content for each user at runtime. Echoing a term used by other companies, he described this as “a segment of one.”
The mixed-and-matched newsletter is thus personalized with the kind of content that Boomtrain platform predicts you will find most relevant — a link to an article that matches your interests, say, or a product image/description that relates to something you once bought.
Gore said the most important data for defining the final newsletter is user activity, such as which stories the email recipient viewed on a website. The Guardian newspaper, for instance, is a client company that has some Boomtrain code snippets on its website to track which stories the user viewed.
Or the data could reflect social sharing of stories, products the user has viewed or bought or other direct or inferred indicators about what the user is interested in, obtained from an Oracle profile, Oracle’s BlueKai data management platform or other sources.
Aside from saving time for the marketer, Gore said that his company has found its platform can lift click-to-open rates by 200 to 300 percent.
While machine learning has been used for personalization in publishing for a while (e.g., Netflix), he noted that it’s often been custom-built rather than available as a service.
He acknowledged there are a number of machine learning-based personalization offerings, including Persado for email marketing and RichRelevance for ecommerce — except they’re not integrated with Oracle.
Opinions expressed in this article are those of the guest author and not necessarily MarTech. Staff authors are listed here.
Related stories
About the author
