Starbucks exec: Balance rigor and flexibility when testing new technology
During the “Omnianalytics: Knowledge Is Good, Now How Can It Transform My Business?” session, executives from Starbucks and Cisco talked about the importance of omnianalytics and leveraging technology to measure the success of new programs. The companies use all sorts of methods such as video and heat graphs to obtain data about new programs in-store.
“You always have to have flexibility in any process that you do and gut feelings are really great, but there are also processes that you can put in place to make sure it’s also rigorous because even if you think it’s the best thing in the market you want to test it to make sure,” said John Goedert, manager of operations testing and innovation at Starbucks Coffee Company, Seattle.
“With all this Big Data that’s coming in, it’s very easy to make the results come out the way you want it, and if you’re not transparent along the way, you may be telling the wrong kind of story,” he said.
“Not being afraid to fail is something you should get very comfortable with if you want to have a robust system in place.”
Starbucks is one of the leaders across verticals when it comes to mobile innovation, with its loyalty and payments offerings.
According to Mr. Goedert, testing new programs like these are very important to Starbucks. Five years ago the company started a Testing and Innovations team, and this team ensures that Starbucks is enhancing the consumer experience on a constant basis.
The coffee retailer uses a three-phase process when rolling out new programs. The first phase is a small pilot test where Starbucks decides whether or not it wants to even pursue the program and continue testing.
The second phase is where the program gets rolled out on a larger scale for a complete set of analytics, and the third phase is the launch validation that tests the program exactly as it would show up when fully launched.
Starbucks uses technology such as FlexSim to analyze how new programs impact in-store service, wait time for consumers and other components of the consumer experience.
One of the main contributors to in-store data for Starbucks is video. Before video, Starbucks would place engineers in-store with a stopwatch to time new programs, but the mere presence of a person holding stopwatch can ruin data since baristas and consumers will notice something is different and may act abnormally.
So much data
When it comes to analyzing data, however, Mr. Goedert acknowledges that there is a surplus of methods for measurement and analysis.
“Decide what methodology you want to do and stick with them,” he said. “There’s many different ways to analyze data and pull it together, aligning on which one you want to use and sticking with it is the key component.
“There’s data for everything, you can find it everywhere but you want to align on how you’re going to use it and be consistent.”
As marketers and retailers are beginning to shift from mobile-first to omnichannel, it is more critical than ever to be able to analyze data from different channels and create a holistic picture.
The traditional abbreviation for marketing channels has been SoLoMo, or social, local and mobile, but Kathryn Howe, retail industry senior advisor at Cisco Systems, Inc., San Jose, adds one more to the mix — photos. Attacking SoLoMoPho requires a lot of effort in terms of truly understanding the customer, but it is crucial to combine all of these channels for an integrated approach.
“We look at omnichannel analytics in layers,” Ms. Howe said. “They start anywhere from a call center, the dot com, the social, the mobile, the in-store experience. Everything is connected to what does it drive in terms of transactions.
“In retail, the driver for information is really our constant need to understand what the shopper is doing and how they make their decision,” she said. “It’s becoming more and more difficult to understand why a shopper decides between you and a competitor because they’re so similar.
“Differentiation requires innovation, and in order to innovate you really need insights. The question is which pieces of insight do you need.”