Infer knows best: Proving the power of predictive analytics by delivering Glengarry leads
The movie “Glengarry Glen Ross” is famous for its desperate salespeople who fight over access to the premium Glengarry leads, knowing they’ll make the difference between success and failure. “These are the Glengarry leads. And to you they're gold, and you don't get them,” Alec Baldwin’s character taunts at one point. This scene may be a work of fiction, but a similar scenario plays out daily inside sales organizations around the world, although with far less transparency.
Leads don’t come in from marketing departments with handy ratings or categorizations that say, this lead has a 75 percent likelihood of turning into a sale, but this lead has only a 15 percent chance. Instead, all leads appear equal at the start, and thus demand time from salespeople to qualify. Salespeople would love to receive only top-performing leads, but that’s never been possible before. Big data and predictive analytics aim to change this narrative.
Infer is Palo Alto startup that pioneered the use of predictive analytics around lead scoring. Now three years into its product roadmap, the company appears to be delivering the kind of results that can meaningfully improve sales performance. Infer, which counts industry leaders like Box, Jive, New Relic, Tableau, Xactly, and Zendesk as clients, doubled its revenue during the period from April through October 2013, and doubled it again in Q4 of this past year, according to founder and CEO Vik Singh. The company raised $10 million in Series A funding from Redpoint Ventures, Andreessen Horowitz*, The Social+Capital Partnership, Sutter Hill Ventures, and Nexus Venture Partners in April of this year.
Singh and his co-founders, who collectively worked at Google, Microsoft, and Yahoo, learned the value of data science in consumer-Web roles, not in the enterprise. But it was the difference in data science adoption between the two markets that led them to focus the efforts of their new company on sales and marketing.
“We were shocked to find out that more data was being applied in the consumer space to enterprise,” Singh Says. “There was very little scientific rigor being applied to decision making by companies.”
Infer look beyond basic signals like what industry a company is in and what channel the lead comes from to evaluate more abstract, and non-obvious signals. This can include basic business metrics like employee count, revenue, geography, ownership structure, trademarks, and patent filings, as well as software use including marketing platform, payment system corporate structure, CMS, analytics platform, email service, and operating systems. The company then layers in social data like job postings, news coverage, social presence, website traffic, advertising sophistication, ecommerce activity, SEO tags, and spam detection.
All of this data is the analyzed, cleansed, and run through Infer’s predictive algorithms which considers historical data to churn out an “Infer Score” – a zero to 100 figure that reflects a leed’s likelihood to turn into business.
“It took us more than two years to build models that would scale with all these data inputs,” Singh says. “We think of this a little bit like ‘Sales and Marketing 3.0.' We take data that you’re sitting on and doing nothing with – maybe you’re using it to create reports about ‘revenue generated this quarter’ – and we use it to predict which leads will drive profit to your bottom line.”
With this score in hand, Infer’s clients can effectively rank all existing leads as well as all lead sources, optimizing the time of their sales teams around selling to top prospects and the spending of their marketing departments around the generation of new hot leads.
“We enable our clients to instantaneously score the leads coming in from their respective marketing campaigns so that they can decide whether it’s good spend,” Singh says. “Usually it takes months to calculate ROI on marketing and, even then, is fuzzy at best.”
One Infer client testimonial from Webroots claims, "Our top scoring leads convert at 6x the baseline. Infer helps us accurately predict winners." Box similarly reports that, "Infer doubled the conversion rate from our highest volume lead sources and uncovered amazing deals that were stuck in nurture.""Across our customers, our predictive models are scoring up leads performing 100 percent to 600 percent better than existing conversion or win rates," Singh says.
The other problem infer aims to solve is something called “nurture,” or when sales teams determining that leads are not ready to buy, for whatever reason, and set them aside for future prospecting. By using Infer scores, salespeople can revisit only those leads that are likely to generate meaningful return on time invested.
“What nurture really means is neglect,” Singh says. “Salespeople have a recency bias. We helped one of our customers close their largest deal of the year from their nurture pile.”
Predictive analytics is still a relatively new category, according to Infer’s founder, who says that very few companies are are turning to this type of technology today with relation to their customer data. As the category grows in prominence, Infer is sure to gain competition. But today, it’s largely greenfield territory, and the company’s biggest challenge is simply educating prospective customers about the power of data science.
“In first year, it was us against their existing process [rather than another competitive service],” Singh says. “Typically, companies don’t even do manual, in-house lead scoring, or if they do, it’s with some sort of aspirational formula based on industry and job title. In these cases, we’ll go in and do a bake-off of our solution versus theirs and compare the results.”
Sales and marketing will always be equal parts art and science. But companies like Infer promise to take away much of the guesswork and allow talented people to ply their craft more efficiently.
(* Disclosure: Andreessen Horowitz partners Marc Andreessen, Jeff Jordan, and Chris Dixon are personal investors in PandoDaily.)