True to brand reputation, Etsy's (NASDAQ:ETSY) IT infrastructure was crafted in-house for years. In fact, the company once showed off its homegrown tech infrastructure -- 3,000 servers kept the site running in the days before Etsy's April 2015 IPO -- and dismissed cloud computing as insufficiently responsive to its needs.
In subsequent years, investors were saying the same thing of Etsy. CEO Josh Silverman has since begun turning things around for the company, focusing on streamlining operations, firing 8% of the workforce, and canceling in-house projects like craft tutorial site Etsy Studio. He's also reassessed how Etsy's IT infrastructure can support the online retailer as big data shapes e-commerce.
Embrace the machine
Machine learning -- i.e. using different high-speed computing tools to analyze pools of data so large that they require computing power beyond the usual spreadsheet or relational database management system -- is emerging as a competitive advantage for large online retailers. Machine learning can suggest related items shoppers might like to buy, customize a website to reflect a customer's demographic profile or shopping history, or help a retailer optimize its pricing strategy.
Another appeal to analyzing big data: machine learning's predictive capacities, i.e. its ability to forecast what shoppers will do at certain times of the year or in response to certain events. Target (NYSE:TGT) provides one of the most notorious examples of big data's predictive capacities getting ahead of the customer -- it correctly sussed out a teen customer's pregnancy before her parents did -- but machine learning boosted its revenue by 15% to 30%. Recommendations generated by machine learning and personalized for each shopper are responsible for 55% of Amazon's (NASDAQ:AMZN) sales.
Etsy's already got recommendations engines in place -- go to the front page of the site as a registered user and you'll see the products the site thinks you'll like -- but CTO Mike Fisher said the company's aiming to boost its use of machine learning tools as it pursues its stated focus of boosting site sales.
The company cited an example of that in its recent earnings call, talking about how machine learning has refined search results by context-specific ranking, offering a user results that take into account factors like the user's location, age, recent activity on Etsy and history of favorites.
Swapping DIY Servers for the Cloud
Machine learning takes considerable computing resources -- and here's where the cloud comes back into play. Cloud services offer their customers access to the computing power necessary for big data crunching. In December 2017, Etsy announced they were moving to Google (NASDAQ:GOOG)(NASDAQ:GOOGL) Cloud. The company's spin:
The migration to Google Cloud comes as Etsy sharpens its focus on growing the core marketplace, prioritizing the buyer and seller experience and increasing the pace of launching new features. Etsy expects that the transition to the cloud will support the execution of its business strategy and will give users an overall better experience worldwide.
The side benefits: Moving to the cloud can reduce overhead costs in-house -- a priority for the company, which has reduced its workforce by 8% and reigned in spending.
And in the most recent quarterly results, Etsy outlined its tech priorities:
In the first quarter of 2018, we invested resources in foundational work, which addressed three primary areas: technical debt, operational efficiency, and infrastructure imperatives. We streamlined our code base to enable more nimble development, implemented a new help center, and made strides migrating to the cloud.
This is Etsy signaling to investors -- and to shoppers -- that it's going to be harnessing the machine-learning advantages that come with Google Cloud and using them to enhance its two primary revenue streams (sales and seller services).
The remainder of 2018 will likely see the company continue to highlight how it's able to boost sales thanks to its new big data focus.