Mercado Libre Goes Big(Query), Plus Cloud Moneyball and Plugging a $300B Retail Search Black Hole
In this issue of Cloudnomics…Mercado Libre uses BigQuery + Looker to score BigWins in e-commerce… How Google Cloud Platform is changing the (Baseball) game… Plus, a new tool that could help retailers plug a $300 Billion(!) a year search abandonment hole.
Mercado Libre Goes Big (Query)
Mercado Libre is obsessed with data.
They have to be.
The Argentine e-commerce giant’s operations make use of massive amounts of data to derive insights from consumer behavior and improve the customer experience, and given that they serve over 65+ Million customers, this is no small feat.
As Jorge Vidaurre, Technical Leader, Data & Analytics at Mercado Libre, writes:
We attempt every effort to make decisions based on data. To infuse our decision making with data, we have found that data needs to be timely, credible, and available for analysis, no matter the source. This includes streaming, collecting and presenting data from our whole ecosystem, which includes external data, our internal management systems, web traffic from products like Google Analytics and App Annie, warehouse and network logs, cloud usage and costs, and, of course, all of our APIs.
Mercado Libre uses BigQuery to process hundreds of terabytes of data and run hundreds of thousands of daily queries to provide the fuel for data-driven insights across the organization.
While BigQuery provides the raw fuel, Mercado Libre turns to the Business Intelligence tool Looker to really kick their insights engine into overdrive.
With BigQuery + Looker, Mercado Libre is building a continuous intelligence system that allows them to feed real-time data to their business, transport and operations teams to inform decision-making.
The result?
Increased speed and adaptability in an incredibly competitive industry, plus the ability to monitor the delivery promise of their shipments to customers and optimize scheduling based on the reliability of their aircraft.
Even better is that this BigQuery + Looker data stack can be integrated with Slack, email notifications, and (for eligible Looker customers) Google Sheets, allowing Mercado Libre’s teams to bring powerful data analysis capabilities into their existing workflows.
Moneyball, Meet The Cloud
When you ask most people to envision the bleeding edge of technology, they’re probably more likely to imagine Elon Musk sweating over the latest SpaceX rocket launch, or delivery drones zipping through neighborhoods, carrying 40-lb bags of cat litter over crowds of nervous pedestrians, or perhaps GPT-3 writing the Great American Novel in milliseconds…
…What they probably are not going to envision is baseball.
And yet, the practice of Sabermetrics (applying statistical analysis to baseball performance — popularized by Michael Lewis’ 2003 book Moneyball, and even more popularized by the 2011 film of the same name) is nothing new. What is new is that the cloud enables a level of Sabermetric analysis that is truly game-changing, with the MLB and Google Cloud leading the way:
From the second batting practice begins to the time a walk-off hit ends the game, MLB is collecting data on the field. Statcast player and ball tracking technology allows for collection and analysis of a massive amount of baseball data in ways that were never possible in the past.
On-field cameras track every movement of the ball and players at 30 frames per second, and that data is fed through Anthos and a Google Kubernetes Engine cluster, then transmitted to the scoreboard and broadcasters.
This live game data can then be fed through the MLB Gameday Engine, combining the live tracking statistics with traditional statistics and providing context for player performances (ex: instantly determining if that 30 ft/sec third base steal was in the top echelon for runners.)
Everything — live, historical and in between — is fed into the MLB Stats API that populates consumer-facing tools like Baseball Savant, where fans can search for things like hit distance and launch angle. It also powers real-time use cases for broadcasters, as well as the MLB app and Film Room. “We’re pulling in data from the API for everything from reviewing major league on-field performance, to player acquisitions, to running our models, to how player performance is going. It’s endless,” said John Krazit, director of baseball systems at the Arizona Diamondbacks™.
This year, the MLB plans to use Hawk-Eye on-field pose tracking data stored in Bigtable to enhance their FieldVision tool, giving fans the ability to see replays from any position on the field.
So, the future of baseball isn’t shaping up to be quite as awesome as teams of Terminator endoskeletons playing baseball in the Apocalyptic ruins of Yankee Stadium, but it’s still shaping up to be really cool (and probably safer!)
Plugging A $300B Retail Search Black Hole
Search abandonment costs retailers in the US $300 Billion a year. Google is looking to change that with the introduction of their new Retail Search tool. According to a recent article on the Google Cloud blog:
Some 94% of U.S. consumers abandoned a shopping session because they received irrelevant search results, according to a 2021 survey conducted by The Harris Poll and Google Cloud. It’s a phenomenon known as “search abandonment.” Indeed, poor product discovery experiences can stop a purchase in its tracks and leave shoppers frustrated. Retailers miss out on a staggering $300 billion each year due to search abandonment in the U.S. alone.
Retail Search allows retailers to add Google-quality search capabilities to their websites and apps. This makes it easier for shoppers to find the exact item they’re looking for, without needing to enter a perfectly-worded query.
Early adopters of Retail Search, like Lowe’s, Fnac Darty, and Casas Pernambucanas are seeing increased conversions, basket sizes, and customer engagement:
Now, through the power of Retail Search, when a shopper searches for a “long black dress with short sleeves and comfortable fit” on an ecommerce site, they should immediately get results for precisely that — rather than refining their search multiple times, or worse, giving up their shopping journey.
Retail Search is a fully-managed Google Cloud solution now generally available to retail customers. You can learn more about it here.
IKEA Gets A 2% Global AOV Boost With The Help of Google’s AI
Albert Bertilsson, Head of Engineering at IKEA, described how the company uses Google Cloud’s Recommendations AI models to match customers with more relevant product recommendations, creating a better experience for the customer and increased AOV (Average Order Value) for IKEA:
Recommendations AI models like ‘Recommended for you’, ‘Frequently Bought Together’ and ‘Others you may like’; are coupled with business goals like optimizing for conversion rate, click through rate and revenue. We experimented with many different model combinations and custom rules. All this was easily configurable right in the GCP console. One of the simplest custom configurations we used was to only recommend items that were in stock, and when items were out of stock we looked at similar items that were available to augment the experience.
These more personalized recommendations allowed IKEA to increase the number of recommendations on a page by 400%, see a 30% improvement in click-through rates, and an increase in global Average Order Value of 2%, which is a heckuva lot of Trollbergets sold!
In the future, IKEA is looking to experiment with using Google’s Vision API Product Search to allow customers to search for products by matching their own product images to sets of similar IKEA products.
If you enjoyed reading this, stay tuned to Cloudnomics for more on how real-world organizations are using The Google Cloud Platform and Machine Learning tools to achieve impressive results.