One of my career highlights was a predictive analytics project I worked on in partnership with the big data analytics firm MAX451.19
Not that long ago, a company would put out the same offer to all of its customers. For example, PetSmart would email their entire database an offer for 25% off dog food. The problem is that some of their customers had cats, fish, or frogs. Cat people considered the offer irrelevant and opted out from further messages. They might even get irritated with the company: “You know I have a cat; why do you keep sending me dog offers?”
Over time, companies got smarter and began segmenting customers. I'm guessing “pet_type” is a field in PetSmart's customer database, and now only dog people get dog offers. This is a good start, but it falls apart for big-ticket items. Once you decide to make a large purchase, like a new laptop, your social media and inbox will magically begin filling up with laptop offers. Are they listening in on your conversations, or reading your mind? That's a topic for another time, but you're searching, and they know it.
You do your research, make your decision, and buy that laptop. What happens next? You continue to get laptop offers for months after the purchase. Unlike dog food, you only buy a computer once every few years, so the ads are ineffective. The company needs to know what you will buy next.
At Pier 1 Imports, when a customer bought a dining table, we didn't want to continue pushing dining tables at that customer. So, what's next? Perhaps it's dinnerware to serve on top of the new table. Maybe not; maybe the table was part of a dining room project, and now the customer is searching for a hutch or a side table. Who knows? I'm just guessing, and so were we.
This was the challenge I took to MAX451 CEO Kristian Kimbro Rickard. Rickard explained that this was the perfect business problem for machine learning. With the right model and enough data, the algorithm can predict the next purchase with better accuracy than legacy algorithms. At the time, I believed the only way to process this much data was to build your own Hadoop environment. This meant onsite servers, storage, and cobbling together open source software. A big data project of this magnitude was beyond the appetite and budget of a home furnishing retailer.
As a boutique firm, on the leading—some say bleeding—edge of big data and machine learning, Rickard knew that Microsoft was building big data and machine learning capabilities into its Azure cloud. As an expert, she had the connections and clout to get our project into a private beta program with Microsoft. This is not something I could have done on my own. My local Microsoft team wasn't even aware of the corporate beta program.
MAX451 brokered the arrangement. Microsoft would provide the environment and data scientists; Pier 1 Imports would provide the data and the business expertise; and MAX451 would manage the project, translate between the business and technical experts, and keep everyone on track. Microsoft was on a strict timeline, requiring a successful case study before its Azure big data product launched. We moved quickly, with tangible results in only six weeks. In a short period, and with a small investment, Pier 1 acquired the ability to send more relevant offers to its customers.
In projects like these, everyone is a winner. Pier 1 came away with a groundbreaking solution, MAX451 expanded its reach and credibility, and Microsoft got its case study20. This story went viral: Microsoft published case study videos featuring Pier 1 executives and our partner, MAX451, walking through the personalized marketing use case scenario and resulting analytics solution,21, 22 and articles in Information Week,23 CIO.com,24, 25 Chain Store Age,26 Retail TouchPoints,27 as well as other retail and technology publications quickly followed.28 Although this was back in 2014, elements of this end-to-end data and analytics solution is still considered groundbreaking today.
Whether you're working with a multibillion-dollar consulting firm or the smallest boutique, the key is finding the right people. Kristian Kimbro Rickard took the time to understand our business needs before she recommended a solution.
Too often, people show up with technical answers, searching for a business problem. I'm often asked, “What's your cloud strategy?” At Vitamin Shoppe, we have a business strategy that we will execute using the cloud. That's not a cloud strategy; it's a health and wellness strategy. The difference here is not subtle. As CIOs, we need to solve business problems. Finding partners who ask the right questions before they offer a solution will put you on the right path to accomplishing that goal.
People are key to your success. In the next chapter, we'll explore how to build a robust people network.
http://doyenne360.com).https://docs.microsoft.com/en-us/archive/blogs/machinelearning/how-azure-ml-partners-are-innovating-for-their-customers.https://youtu.be/dJMZVaBWBH4.https://youtu.be/CdYvPgXc5ic.https://www.informationweek.com/big-data-analytics/microsoft-azure-machine-learning-pier-1-digs-in.https://www.cio.com/article/247108/the-cloud-s-game-changer-is-competitive-advantage.html.https://www.cio.com/article/230488/12-microsoft-power-bi-success-stories.html.https://chainstoreage.com/exclusive-content/pier-1-case-study.https://www.retailtouchpoints.com/features/retail-success-stories/pier-1-imports-redesigns-business-intelligence-with-microsoft.https://www.securityroundtable.org.