Versium Launches Self-Service Predictive Lead Scoring With Automated Modeling
Redmond, Washington-based company says this service, utilizing its own billion-attributes consumer data layer, is the first of its kind.
Predictive lead scoring has become a common service for many marketers who want a better idea about which potential customers and current customers are good bets for future sales.
Today, data tech firm Versium launches what it describes as the first self-service predictive lead scoring that generates fine-tuned models built around machine learning and a large consumer data layer.
Called the LifeData Predictive Lead Score service and available through the company’s Datafinder site, the solution employs machine learning to build granular models around a brand’s historical customer profiles, and it then uses those scores to rate lists of potential B2C prospects.
CEO Chris Matty told me a key differentiator is that while there are automated rule-based predictive lead scoring systems, the new LifeData service automatically builds a customized predictive model.
While Versium had previously created predictive lead scoring on a custom project basis for clients, this is its first publicly available service for such rating. The Datafinder site also offers Versium’s online services for data cleaning, validating and other data management.
In a rule-based approach, a brand might discover that many of its best customers are, say, mothers of small children who have SUVs and live on the West Coast. So it applies this rule — find mothers of small children who have SUVs and live on the West Coast — to profiles of potential prospects it obtains. The same marketing gets sent to everyone who meets the conditions of that rule.
A model-based approach, Matty pointed out, is like a rule-based approach, but applied to smaller groups of customers or to individuals. That is, it’s a rule of 1.
In Versium’s new service, a brand uploads info on customers who made purchases in one or more of its campaigns, as well as — if it can — profiles of those who were similarly pitched but didn’t buy. Obviously, the more info on each person the brand can supply, the better.
Versium then enhances the brand’s profiles with its massive LifeData data layer. Matty said that its data layer currently contains more than a trillion customer and business attributes covering online behavior, social behavior, purchase interest, financial data, demographics and other types.
LifeData, he said, is compiled from three sources, each representing about a third of the total. There’s commercially licensed third-party data, public records like census info or divorce records and data that the company has indexed via its own crawlers. This latter category includes social networks, location data and weather records.
Within several hours of uploading a brand’s historical customer data, he said, the new automated service will generate predictive profile models for small groups of customers or individuals. This can then be applied to other lists of prospective customers in order to score those that might be better bets for targeting, based on the models of which customers responded before and which ones didn’t.
A slider control also allows the marketer to select the bottom threshold of scoring, so that marketing efforts can be directed only at the most likely future customers.
The service, which requires no technical support, can then export the scored leads via an API to a customer relationship management system or other marketing platform for campaign implementation.
This kind of focused targeting, Matty claims, can reduce marketing costs by as much as 40 percent because resources are spent only on the best bets.
The company also said the LifeData predictive service can boost conversion rates 400 to 500 percent and claimed that in at least one case, it boosted a brand’s conversion by 900 percent. The service runs $4,000 per month, and a brand can generate up to 100 models in whatever time frame, up to a year.
Opinions expressed in this article are those of the guest author and not necessarily MarTech Today. Staff authors are listed here.