Salt Lake City, Utah, and the surrounding area are home to an under-appreciated number of technology companies, including one that’s using an immense amount of data to help salespeople put their energies in the right places. It’s not exactly saving the world, but if the technology works as advertised, it’s actually a rather impressive display of data analysis.
Provo, Utah–based InsideSales, which launched in 2004, uses a long list of big data technologies and techniques in order to figure out which sales leads are the most likely to buy and, therefore, which ones are actually worth a salesperson’s attention on any given day. The company is the brainchild of its founder and CEO Dave Elkington, who describes its strategy as essentially trying to “algorithmically execute my philosophical approach” that like-minded people are likely to behave in similar ways. And the more you can model those behavior patterns, down to very specific sets of circumstances, the better you can predict what they’ll do.
The culmination of this strategy came in November 2013, when the company launched its Neuralytics platform, which Elkington calls “a massive context engine.” It aggregates data from its users’ 150 million profiles of their customers, and more than 10 billion sales interactions with those people. Every new interaction or new customer means more data points for the system to learn from. There’s lots of external data to analyze, too — everything from weather data to sports scores. (According to Elkington, there’s a 72-hour window after a big win when sports fans are more likely to buy.)
Some of it sits in a large MongoDB cluster, while other stuff resides in Hadoop and HBase. On top of it all is a graph database that maps all of these characteristics and buying behaviors. Elkington said his team of engineers and statisticians use whatever tools are at their disposal — neural networks, support vector machines, Bayesian statistics, whatever — to build models that can best predict someone’s buying preferences and assign them a score and profile.
InsideSales’ weights its algorithms “very, very” granularly, he explained, so that “every situation has a learning set” that’s constantly being retrained. It’s not enough to just profile every lead and say “Leads that fit this profile act like this.” Elkington wants his system to know how every little variable, including the personality profile of the salesperson, is going to affect someone’s behavior.
The company has a whole suite of products that use these models, including one that analyzes phone calls and provide sales reps real-time feedback about how a call is going and how they can close the deal or, if need be, save the call. Elkington said the company has an email product that does something similar, analyzing the text of messages and scoring them based on the likelihood the recipient will interact with it. It will suggest rewrites, and even schedule emails to send at a time the system has determined the person is most likely to open it and respond.
“Frankly,” he said, “it’s pretty badass.”
On Monday, the company plans to announce a new product called NeuralView that takes all of the scoring and modeling it’s doing and actually prescribes the best ways to interact with a lead. This means telling customers not only how likely someone is to buy, but also when and how to contact them, and possibly which salesperson should contact them. Elkington acknowledges InsideSales isn’t alone in doing predictive scoring for sales leads — startups such as Infer and Lattice engines are also doing it — but he says this broad prescriptive information is something no one else is yet offering.
Talking to Elkington, it’s easy to hear how excited he is about the technology, and it might be easy to dismiss him as just another CEO trying to sell a vision that’s too good to be true. Indeed, the sheer scale of the data analysis, and the level of granularity it tries to reach, could be criticized as overkill if not far-fetched.
But if you’re inclined to put faith in numbers, it appears InsideSales is doing something right. The company recently closed a $100 million infusion of venture capital — some of which came from CRM kingpin Salesforce.com — and has a customer list of 2,000 and growing. Elkington said that list, and his employee count, are growing by about 100 per month.