1plusX adds search-like keyword targeting to its DMP
The Zurich-based company employs machine learning to predict which consumers might be appropriate for a given set of terms.
The first commercial customer signed up this week for a data management platform’s new keyword-based, “search engine-like” targeting.
European ad network Goldbach Group is the first to offer keyword targeting for brands’ campaigns through the DMP 1plusX.
Recently developed by the Zurich-based firm, the keyword targeting lets advertisers find the audiences they want to target online using any word or word combination, instead of the predefined or custom combined segments often offered by other DMPs.
1plusX Head of Product Roger Gatti suggested to me the example of an advertiser who types in “skiing,” “skating,” and even the names of some winter resorts, when the intent is to reach consumers interested in the topic of “winter sports.” The advertiser can also exclude results for “Korea” or “Olympics” if they want to avoid consumers whose winter sports interest is primarily oriented around the 2018 Winter Olympics.
Or the advertiser could type in any set of words. The platform’s machine learning judges whether to add attributes to consumer profiles based on cookies, interaction with content and other data. Here’s a screen shot, showing the number of visitors to a given URL whose attributes match the keywords:
When an advertiser enters a new keyword, the platform assesses the probability that consumers are interested in that topic, and the advertiser can control the probability of the match. For instance: show me consumers who are more than 80 percent likely to be interested in, say, beach weddings.
The probability score, Gatti said, allows brands to determine the tradeoff between quality and reach for every profile attribute. He said the standard number of attributes for each profile is about 10, but keyword inquiries can lead the platform to begin building new attributes. Most of the data in 1plusX is first-party, and the profiles are generally siloed by client.
The platform is primarily used for ad targeting, he added, but it is also being employed for content recommendations. In addition to targeting existing audiences, it can be utilized to create new audiences.
Gatti said that involves more than typical “lookalike” matching, where new audiences are generated because these potential new customers have the same attributes as those existing customers. Instead, he said, the platform probabilistically matches whether these new potential customers might have these same interests.
Opinions expressed in this article are those of the guest author and not necessarily MarTech. Staff authors are listed here.