Miam: Serving up smart recipes with AI powered by Google Cloud
About Miam
Miam makes sure shoppers always have the right ingredients at home to cook the recipes they love. By suggesting recipes uniquely tailored to each shopper and automatically adding all necessary ingredients to their basket, Miam’s white-label solution helps online grocery retailers boost sales while delivering an easier, faster, and more intuitive shopping experience.
Tell us your challenge. We're here to help.
Contact usMiam launches a white-label solution for retailers on App Engine, scaling from 100 to 30,000 daily users while delivering a 40% increase in the average basket value of its first retail client.
Google Cloud results
- Lets eight-person team focus on product development by automating deployments with App Engine
- Boosts retail client’s sales, delivering 40% increase in average basket value with high accuracy
- Accelerates growth by delivering performance at scale, with option to switch to autoscaling in App Engine
- Improves user experience by increasing accuracy of converting ingredients into products from 80% to 97%
Seamlessly scaling from 100 to 30,000 daily users
Have you ever been in the middle of cooking a new recipe, just to realize that you’re missing a key ingredient or two? If you find it difficult to plan your meals and keep track of what’s in the fridge, you’re not alone. César Tonnoir and his colleagues found it difficult, too, and decided to do something about it. That’s how Miam was born.
“We love food and enjoy some of our most wonderful moments around the table with our friends and families, but we don’t like to spend our time shopping for groceries every week,” says Tonnoir, CEO and co-founder of Miam. “So we tried to create a solution to help people avoid this task and still get all the groceries they need.”
Miam makes grocery shopping easier, faster, and more intuitive with AI. By integrating Miam’s white-label solution seamlessly into their websites, retailers can boost sales while delivering a better customer experience to shoppers. Miam suggests recipes based on their shopping preferences and adds all the necessary ingredients to the basket with just a click.
“We’re a small team of eight people, so we don’t want anyone to spend their time maintaining our infrastructure. From day one, we decided to fully automate all deployments and testing processes, and needed a technology partner to help us achieve that. That’s why we partnered with Google Cloud.”
—César Tonnoir, CEO and co-founder, MiamMore food, less fuss
Miam helps food lovers focus on their passion by automating the rest of the process, and the company’s approach to technology mirrors that philosophy. “We’re a small team of eight people, so we don’t want anyone to spend their time maintaining our infrastructure,” says Tonnoir. “From day one, we decided to fully automate all deployments and testing processes, and needed a technology partner to help us achieve that. That’s why we partnered with Google Cloud.”
Tonnoir and his colleagues had worked with other cloud solutions in the past, but hadn’t found the perfect combination of performance and usability. To wow audiences with prototypes early, fast deployments were initially crucial, as were low costs, because the small startup had few cash reserves. But looking beyond the early days, Miam’s ambitious growth plans required a solution that could also scale with resilience. When Tonnoir tried Google Cloud for the first time, he was impressed with its user friendliness. “I was really amazed by the simplicity of Google Cloud, and thought if I could learn the basics so easily, so could my team,” says Tonnoir. “Google Cloud delivered the combination of speed and agility we needed, coupled with a startup program that tailored to our needs as a young company, for example with free credits.”
Dishing up simplicity with App Engine
Today, Miam’s white label retail solution mainly runs on Google App Engine, consisting of three main components. An API, written with Ruby on Rails, acts as the interface between all retail customers and Miam’s features and algorithms. A Python app runs the AI algorithms themselves, and an SDK library built with Angular can be deployed on any website. Databases are managed with Cloud SQL.
What the Miam team loves most about working with this setup is the simplicity. “App Engine is easy to configure with a simple YAML file, and on the projects so far deployed, the configuration file hasn’t even exceeded 40 lines of code,” says Tonnoir. “That’s enough to run a full application, but requires very little effort to maintain and adapt.”
With different projects running on App Engine concurrently, its easily customizable infrastructure also helps the Miam team manage multiple stacks concurrently. It’s simple, for example, to build and run different containers for Ruby or Python. “Because of this ability to build containers for each stack, we don’t need to bootstrap our own Docker container, we can just use the one provided by App Engine,” says Tonnoir. “It runs with barely any maintenance from us.”
Going from 100 to 30,000 daily users
With its solution up and running on App Engine, Miam was ready to enhance the shopping experience of retailers around the country. The first moment of truth came in February 2021, when Miam went live on French ecommerce and shopping website Cora.fr. “We needed to be able to guarantee our client that nothing would crash when they activated the service,” says Tonnoir. “And that’s what we achieved with Google Cloud.”
To ensure resilience, the Miam team load tested its solution prior to the launch on Cora.fr, scrutinizing the latency and load of the application. “With monitoring capabilities and detailed charts, App Engine lets us scale our infrastructure perfectly, and we knew exactly how many servers we needed to add,” says Tonnoir. “From one day to the next, we went from 100 daily users to 30,000 without any hiccups and have been running smoothly ever since.”
So far, the Miam team is scaling manually, but as the company grows, Tonnoir believes that the automatic scaling features of App Engine could become crucial. “It’s reassuring to know that we can switch from manual to automatic scaling in a heartbeat,” he says. “It makes me confident that we’ll be able to scale our solution with our growth and bring Miam to more and more retailers without issue.”
“Our standout feature is that customers receive personalized recipes and can buy all those ingredients directly and reliably. By building customized AI algorithms and leveraging Cloud Natural Language AI to analyze text, we were able to increase the accuracy of converting ingredients into products from 80% to 97%.”
—César Tonnoir, CEO and co-founder, MiamFrom ingredients to products with 97% accuracy
At Cora.fr, the implementation of Miam’s white-label solution tool has been seamless. Today, customers see personalized recipe suggestions and can add all the necessary ingredients with a click of a button. By providing this smart and convenient way for shoppers to fill their pantries, Cora.fr has been able to measure an increase of the average basket of more than 40% if the basket contains recipes.
This success hinges on the quality of Miam’s AI algorithms and the precision of converting recipes into full shopping baskets. In this regard, Miam has come a long way: when the team was developing the solution’s prototypes, converting ingredients into actual, purchasable products on retailer’s websites had an accuracy of about 80%. This would have meant missing ingredients in customers’ baskets.
Today, the accuracy is near perfect. “Our standout feature is that customers receive personalized recipes, and can buy all those ingredients directly and reliably,” says Tonnoir. “By building customized AI algorithms and leveraging Cloud Natural Language to analyze text, we were able to increase the accuracy of converting ingredients into products from 80% to 97%.”
“By building our infrastructure on Google Cloud, we’ve been able to fully focus and launch our solution by automating away the rest. With everything running so well, we’re ready to scale. That’s why staying with Google Cloud is a given, and we’d like to expand the partnership in the future.”
—César Tonnoir, CEO and co-founder, MiamA match made in foodie heaven
As major foodies, Tonnoir and the entire Miam team aren’t starved for ideas when it comes to new features. Soon, customers will be able to add their own recipes to their favorite retailers’ websites and easily convert them into full shopping baskets. In the future, the team might use more features of the Vertex AI platform to make its algorithms even more accurate.
Meanwhile, they’re also working on a meal planning feature to help shoppers spend even less time planning their meals, and more time enjoying them. “We actually got the inspiration from Google Calendar,” says Tonnoir. “Customers will be able to just plan their weeks with a meal plan calendar to add their favorite recipes and add all the ingredients for all of them automatically, saving lots of time.”
Giving foodies more time for food is what Miam is all about, and Google Cloud has been an effective test kitchen to bring their ideas to life. “By building our infrastructure on Google Cloud, we’ve been able to fully focus and launch our solution by automating away the rest,” says Tonnoir. “With everything running so well, we’re ready to scale. That’s why staying with Google Cloud is a given, and we’d like to expand the partnership in the future.”
Tell us your challenge. We're here to help.
Contact us
About Miam
Miam makes sure shoppers always have the right ingredients at home to cook the recipes they love. By suggesting recipes uniquely tailored to each shopper and automatically adding all necessary ingredients to their basket, Miam’s white-label solution helps online grocery retailers boost sales while delivering an easier, faster, and more intuitive shopping experience.