WineAdvisor unites 80,000 connoisseurs with Compute Engine
WineAdvisor’s smartphone app helps wine connoisseurs, also known as “oenophiles,” discover and rate new varieties and vintages. The company connects with its rapidly growing user base with real-time interactions powered by Google Cloud Platform.
Using image recognition to rate wine varieties
There are more than four million wine lovers with smartphones in France, and WineAdvisor’s app aims squarely at this ardent group. “There are more people talking about wine than sports in France,” says Cédric Morelle, founder of the nine-month-old startup. “We see it as a huge opportunity. We are the very first company to provide a social and mobile commerce app around this.”
As people post pictures of wine labels, the company’s image recognition software identifies the wine and sends them information about the variety and vintage, including a wine fact sheet, ratings, and reviews from both professionals and the user community. Wineries also benefit, as they can get detailed market data about who is interested in their product, as well as access to a new selling channel — people can order wine directly through the app.
The company already has more than 80,000 active users, with more than 1.8 million recognized vintages. Three-quarters of France’s wineries are already in the database, and Morelle plans to expand the service across Europe.
With a rapidly expanding business based on real-time user interactions, Morelle knew that the company needed a high-performance computing platform that could scale to meet rapid growth. At the same time, the young company needed to keep costs low.
The power of reliability and flexibility
For Morelle, the clear answer was a cloud-based platform. The team benchmarked solutions from all of the major vendors, but found that most choices were bare bones IaaS solutions that lacked features like automated provisioning. “We needed to be able to build a specific architecture at a low price,” he says. “At the same time, with a small team of about 10 people, we didn’t want to waste time on chores like changing servers and migrating data.”
Google Cloud Platform offered Morelle an affordable opportunity to build on an automated platform, and the company jumped on board.
The company’s image recognition system uses several instances of Google Compute Engine running in parallel. It also uses Google BigQuery for statistical market analysis for winery partners. “We can point out buying trends and excellent reviews, which might trigger an online promotion by the winery,” Morelle says.
The system’s solid performance has not faltered, even as the company’s user base expands. “Google Cloud Platform is one of the most reliable systems I’ve ever worked with,” Morelle says. “I’ve tested most of the others out there, such as Amazon, and it’s much more reliable.”
Compute Engine also gives Morelle’s team the flexibility to easily modify the architecture for performance improvement. “It’s very easy to quickly change the configuration of the app and architecture with two or three clicks, which is something that other vendors don’t offer,” he says. “We were able to quickly adjust the way our image recognition algorithm consumed CPUs and RAM to make it much more efficient, for example.”
Ditching pesky IT chores
Morelle’s decision to use Google Cloud Platform lets his team focus on strategy and execution rather than IT chores. The company plans to expand to the U.K. market in the next several months, and the platform’s automated scaling makes provisioning and maintenance chores a non-issue. “Provisioning probably won’t take more than 10 minutes,” he says. “If we did that on-premise, it would take almost the equivalent of a full-time developer to handle such maintenance chores.”
Moreover, user response time continues to drop as the team tweaks its system for peak performance. “We’ve been able to cut user response time by more than 50 percent using Compute Engine,” he says.
As the company continues to expand across countries and mobile platforms, Morelle finds it reassuring to know that its core platform will keep pace.