Better personalization through predictive analytics

If you buy a television on Amazon.com, the next time you visit the site, you’ll see ads for ... televisions. That doesn’t make much sense, does it?

Now alumni-led startup Infinite Analytics has built a better way to match consumers with products and content to personalize a user’s experience. Their platform, which began as a class project, is at the forefront of a big-data trend—predictive analytics.

Infinite Analytics works both sides of the giant digital data pile: customers and products. First, they collect and analyze site information and public information that people have put out about themselves, from customer behavior on the site to social data to macro trends data.

“We disambiguate it, identify the same user across sources, and create a 360-degree profile of the user,” says co-founder Purushotham Botla, SDM ’13. Then, they run the profile through algorithms that tell consumers the most interesting products to them, making the whole product discovery process easier and faster.

This is the kind of thing we’ve grown used to from shopping on sites, like Amazon, that make recommendations based on what you’ve already bought. The difference is, a site running Infinite Analytics can serve up a tailored recommendation to you on your very first visit.

On the product side, Infinite Analytics pores over all the images and text about a company’s products and builds a semantic profile of the products, so products can be categorized by function, occasion, and so on. You probably remember the dress that generated ten million tweets in one week last February. Was it blue and black, or white and gold?

“We were able to break it down: the dress was 80 percent blue, 10 percent gray, and 5 percent yellow. We could tell people that the debate was over,” Botla says. That was just for fun, but the co-founders say their detailed product databases deliver real results. One customer, India’s Cromā Retail, ran A/B testing and saw online conversions jump 217 percent with Infinite Analytics. Typically, customers are reporting a 25 percent boost, still high, the company’s founders say.

Bootstrapping Before enrolling at MIT Sloan, co-founder Akash Bhatia, MBA ’12, founded a Ticketmaster-style startup in India. Botla built trading platforms for Fidelity Investments while he worked his way through the MIT System Design and Management program. The pair met ina class called Linked Data Ventures, taught by World Wide Web inventor Sir Tim Berners-Lee. They created a prototype as a class project. That prototype advanced to the semifinals ofthe 2012 MIT $100K Entrepreneurship Competition, and they joined the inaugural MIT Founders’ Skills Accelerator (now the Global Founders’ Skills Accelerator) cohort the same year. They have raised two rounds of funding from investors in Asia, Australia, Europe, and the U.S.But it was in-kind deals from cloud infrastructure providers such as Microsoft, IBM, and Rackspace that provided the much needed server capital to build an analytics platform in the early stages when funds were very limited.

“That’s equivalent to almost $250,000 in funding,” Botla says. “People don’t realize the value of in-kind donations when you’re bootstrapping a company.” Infinite Analytics now has 15 employees in Boston, Macedonia, and Mumbai, and 17 customers around the world.

A pivot to business The company was originally called Schmooze-Butler because Bhatia and Botla aimed to match people with like-minded peers for hanging out. It was a name that “people really liked or really hated,” Bhatia says, and it made less sense once the pair shifted their focus to e-commerce. On the bus back to Boston froma meeting with a potential customer in New York City, Bhatia and Botla brainstormed the new name, a sly nod to MIT’s Infinite Corridor, as well as a description of the technology’s growth potential.“It has infinite applications,” Bhatia says, adding that custom news and entertainment will probably be the company’s next market.

MIT professor Berners-Lee and Twitter’s chief media scientist Deb Roy, an associate professor at MIT, are advisors to the company, as is MIT SloanProfessor Erik Brynjolfsson, director of the MIT Initiative on the Digital Economy and co-director of the Analytics Lab action learning seminar. Brynjolfsson says Bhatia and Botla are pioneers in an increasingly popular field.

“I’m proud that this is part of a growing ecosystem around analytics that’s centered here around MIT and Kendall Square,” Brynjolfsson says. “They are a great example of how to take some tools we teach at MIT and use them to make the world a better place and provide value for customers.”