Optimizing assortments with big data

Kerem Tomak, Macys.com

Macy's Inc. is one of the nation’s largest and most well-known retailers, with a loyal base of customers who shop at its stores and online. The company’s e-commerce division, Macys.com, is relying on big data and high-performance analytics software from SAS to better understand its customers and help increase overall profitability. Kerem Tomak, Vice President of Analytics for Macys.com, recently shared some insights on this.

Alison Bolen: What opportunities or challenges does big data create for your organization?
Kerem Tomak: Basically being able to understand how the business is performing in the marketplace. For example, it’s about knowing how our products are selling on Macys.com, how we are selling in the stores. What’s the impact of our marketing efforts on the sales results that we are seeing both online and in the stores? The challenge is to gather data and turn it into insights in a timely, almost a daily fashion, so as to respond to any kind of consumer demand changes or marketplace changes. And we need to be able to respond to it fast enough so that it really does make a difference in how the business responds to what we see in the data.

Bolen: How can high-performance analytics help you embrace those opportunities?
Tomak: That is our first frontal attack, to be able to process the data and put it in a form that makes sense not only to us as analytics people or data people, but to the business people, senior management and the like so that they understand the story behind the data. Then they can make decisions based on that information. That is where I believe high-performance analytics is going to really shine and add value to the business.

Bolen: One of the benefits of working with SAS is that you can do logistic regression with big data, not just some of the summary statistics and basic reporting on big data.
Tomak: Right. We have been using SAS by pulling data from Hadoop and passing it into a variety of SAS procedures and modeling. We’re able to do things with these models that we weren’t able to do before. Using SAS in more of a high-performance environment allows us to manage these models and the big data itself in a fast fashion so that we get results. Generating hundreds of thousands of models on granular data versus only 10, 20 or the 100 that we used to be able to run on aggregate data is really the key difference between what we can do now and what we will be able to do with high-performance computing.

Bolen: Are there other ways that high-performance analytics is influencing the competitive environment in your industry?
Tomak: One direct application area is to be able to understand data at the product level. Traditionally we would not analyze sell-through, out of stock or price-promo response at the store and SKU level, but several levels up at the product hierarchy segment level or some higher level. We would aggregate away from the products, and try to extrapolate to understand what products are more readily available for modeling purposes.

But now you can run hundreds and thousands of models at the product level - at the SKU level - because you have the big data and analytics to support those models at that level. It really brings us to where we want to be. “What is the probability of selling a product with a set of attributes at a particular time and location?” is one of a list of questions we can answer with HPA.

That is really the next, deeper level of assortment optimization that we can aim for with high-performance analytics.

Bolen: With high performance can you look at more granular levels for each customer?
Tomak: Yes, for each customer as well as for each product that is being sold. Where is the location of the product? Is it sold through from Macys .com? Is it sold through the store? Is the customer able to find the product at the store? If not, can he or she buy from Macys.com and vice versa?

From what I’m observing recently, there’s also more interest in using high-performance computing to understand display ad performance on the web. Ideally, ads can be shown to those people who will be more willing to see them and click through in the first place. High performance analytics helps increase the relevancy of the ads themselves, looking at all the information that’s out there about the ads, content and information that is conveyed to the browser.

Find more answers to your tough "big data" questions in this special 32-page report on high-performance analytics.

tags: high-performance analytics, macy's, retail

One Trackback

  1. [...] Optimizing assortments with big data and high-performance analytics [...]

Post a Comment

Your email is never published nor shared. Required fields are marked *

*
*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <p> <pre lang="" line="" escaped=""> <q cite=""> <strike> <strong>