(Cross-posted on the Google Cloud Blog.)
Editor's note: Today’s guest blog comes from Devavrat Shah, Chief Scientist and Co-founder of Celect, which helps retailers understand buying patterns and customer choices.
Retailers spend a lot of time and money trying to figure out what people will buy and when, online or offline. Many retailers see this as an art, but at Celect, we want to add science to this process. The answer lies in what we call the “Choice Engine,” which gathers data on what customers buy and don’t buy – instead of just simply finding out how they rate products they like. Think of the shopping process this way: If someone browses black shirts and red shirts online, but puts a blue shirt in the shopping cart, they’re giving you comparative information. Celect can take these choices and suggest which products a retailer should stock more or less of – as well as predict when price becomes a factor in a shopper’s purchase decision.
My cofounder Vivek Farias and I, both professors at MIT, decided to put our brains together and see if we could bring our technology to the commercial market. We knew our technology was great, so we bootstrapped a team together – two professors, two engineers, and one person on the ground doing business development. Our biggest challenge was scaling our technology even though we had an extremely small development team. We didn’t want to run a system when we didn’t yet have clients.
Fortunately for Celect, we met the criteria for Google Cloud Platform for Startups, giving us $100,000 in credit for Google Cloud Platform products and easy access to engineers and architects to help us make the most of our infrastructure. We quickly found out how good Google’s documentation is, which matters when you’re a startup that needs to move quickly. We get to tap into the expertise of people who’ve spent 10 years building cloud infrastructure, and they know it very well. The web user interface of Google Cloud Storage is very intuitive to navigate, and gives us an overall view of the system and the resources in use.
We run our workloads on Google Compute Engine, which operates easily with our commodity Linux machines – another way we save money as a startup. Google Cloud Platform also gives us peace of mind about security. Retailers trust us with highly proprietary information, and they’re very sensitive to data breaches. When they hear we rely on Google, retailers know we’re adhering to strong security standards.
Since we’re going after large retailers for our product, we need the scalability to store massive datasets. We can create new data stores in Google Cloud Platform so that every client’s data is siloed from the others. It’s the perfect on-demand infrastructure for a company like ours that needs to run lean for the first couple of years.
At this stage in our growth, we want to make very efficient use of every dollar we spend. The past year has been very successful for us, with some great retailer brands signed on and a threefold growth in employees. Google Cloud Platform will grow with us, while helping us develop our products better and faster.
- Posted by Devavrat Shah, Chief Scientist and Co-founder, Celect