McCormick & Company, Incorporated is a global leader in the flavor industry with more than $4 billion in annual sales. McCormick manufactures, markets and distributes spices, seasoning mixes, condiments and many other flavorful products to the entire food industry—retail outlets, food manufacturers and food-service businesses—in more than 125 countries and territories. Since Willoughby M. McCormick founded the company selling root beer extract in 1889, McCormick has demonstrated a strong commitment to the communities in which it operates and the planet as a whole. Innovation in flavor and a clear focus on employee engagement and product quality has allowed McCormick to grow its business globally and become the flavor leader it is today.
Earlier this year, McCormick launched FlavorPrint, a tool that combines predictive analysis with McCormick’s deep understanding of the human palate to help its customers put together meals that will delight everyone around their table without relying on old standbys.
FlavorPrint assesses each user’s tastes and palate and gives them a flavor fingerprint, not unlike a Myers-Briggs assessment, and then offers personalized recommendations for recipes, meals and eventually wine pairings. That comes in handy during holidays like Thanksgiving, when a family can take into account every guest’s individual FlavorPrint to come up with a meal that will please all of them. “FlavorPrint is truly predictive,” says Bjoern Leyser, McCormick Vice President, Global Connected Consumer. “FlavorPrint combines sensory science and culinary science with the insights McCormick has been gathering for many, many years.”
It’s an idea whose time has come, as McCormick’s research suggests that most of us are creatures of habit when it comes to the kitchen. “Most people have seven to nine meals that they make over and over again,” Leyser says. “But for whatever reason, breaking out of that rut and making a new dish is really difficult.”
McCormick, a company known for its expertise in natural flavors and sensory science, has been gathering insights about consumer desires and behaviors for decades. The company decided to combine its expertise with cutting-edge technology to solve the dinner dilemma in an easy and truly personal way. The company partnered with Enterra Solutions to develop algorithms matching people’s taste preferences to flavors, and the FlavorPrint predictive recommendation platform was born.
In 2013, McCormick launched a small beta program on a hosting provider. Word spread, and even in beta mode, the service quickly grew to 100,000 participants and produced a sales increase of 4.9 percent for McCormick—an unheard-of spike in a category so mature that even a 1 percent increase in sales is big news. McCormick knew that to accommodate the surge in use, the FlavorPrint experience would have to be as nimble and flexible as possible, so the company began looking for a platform that could scale to handle millions of users around the world. “We knew FlavorPrint would scale exponentially, both in terms of users and recipes—and in real time,” Leyser says. “We needed a solution that could offer us robust data storage, seamless scalability, and low cost."
McCormick chose AWS to run the FlavorPrint service experience, leveraging the low cost and scalability of Amazon CloudFront as a content delivery network (CDN). “AWS was the clear choice over other cloud providers because of their extensive APIs, which enable clear, simple management of resources,” says Samir Rohatgi, Enterra Vice President of Professional Services. “And using Amazon CloudFront as our CDN gave us much more flexibility to learn and grow.”
FlavorPrint uses Enterra’s Cognitive Reasoning Platform (CRP) to create each product and recipes’ specific flavor profile, or FlavorPrint mark. AWS is used to host the FlavorPrint service, and data like food photography, session data, and recipes. Consumers go to the McCormick.com website, launch FlavorPrint, answer some questions about their tastes, and learn about their unique and personal FlavorPrint. After that, they receive personalized, predictive meal recommendations that take into account a variety of factors such as day of week or time of day depending on your habits. (For example, what you cook on the weekend might be different from what you might eat at the end of a long day at work.) The service incorporates health and wellness preferences—like recommending foods that are low in sodium or high in fiber or excluding avoidances like nuts or gluten.
In the two months after the service went live, its traffic has doubled, and the company expects this growth to accelerate as the winter holiday season approaches. McCormick plans to begin partnering with other companies like Amazon Fresh, enabling partners to plug FlavorPrint right into their site so their own customers can begin using it.
The company uses Amazon Elastic Compute Cloud (Amazon EC2 – General Purpose, Compute Optimized and Memory Optimized) to provide FlavorPrint Services to sites like McCormick.com, including the recommendation API and the widgets that partners can plug into their sites. The FlavorPrint platform consists of a Java Application Server API layer that calls business logic in the middle tier, and reads and writes data to MongoDB. The service’s virtual machines run on Linux. Auto Scaling and Elastic Load Balancing are used to keep site traffic running smoothly.
McCormick uses Amazon CloudFront to store each user’s FlavorPrint mark and serve them, as well as recipes and images, to the website. “Our services are consumer-facing and must integrate easily with the client site,” Rohatgi says, “so we need to be hyper-responsive and scalable. AWS enables us to do that.” The site is deployed across multiple Availability Zones.
The company also uses Amazon Simple Storage Service (Amazon S3) to store logs, FlavorPrint marks, images, recipes, and other data. “Amazon S3 works really well with CloudFront for the CDN,” says Dan Slattery, team lead for FlavorPrint Connected Consumer. “We use Amazon S3 and subsequently CloudFront to store static recipe files, which allow us to quickly send out recipe details so a partner can access it as necessary.”
By using AWS, McCormick has doubled processing capacity and still cut infrastructure costs by 50 percent compared to what they were spending with its former hosting provider. “Because we can chose which instance type to use, we can stand up development, QA, and production environments for less than we used to spend on production previously,” Slattery says. “We’re saving a substantial amount of money—we’re cheaper and we’re more scalable than we were before moving to AWS.” Garriott adds, “We can spin up additional servers as needed for testing and development, which makes for much easier platform testing. By using AWS, we can get things up and running fast and grow as we need to. AWS is helping FlavorPrint succeed.”
AWS is not only cost-effective, but its ease of use has helped McCormick get to market quickly. It took McCormick about half a day to stand up its QA environment, for example. “It wouldn’t be possible to move this fast with our former provider,” Slattery says. “Recently, we spun up a new batch process and scaled our environment 10x in a few hours to support a massive data migration effort. AWS is an incredibly powerful infrastructure.” Garriott, says, “We love AWS—it allows us agility in a fast moving environment. AWS earns our business every single hour we use them, and that’s not something we say lightly.”
The company uses AWS to adjust to its rapid growth. “AWS gives us endless scalability,” Slattery says. “It’s just a much more flexible system. We use Auto Scaling, and from an infrastructure standpoint, we can scale appropriately and handle the increased load at the holidays.”
The company's use of AWS enables it to help it continue to build products and services that delight customers around the world. “AWS lets us innovate fast with little or no risk,” Garriott says. “Using AWS seems like such a simple decision, and yet it’s turned into a game-changer for us.”
To learn more about how AWS can help you run your website, visit our Web and Mobile details page: http://aws.amazon.com/web-mobile-social/.