5 Q’s on Data Innovation with Tim Callan
Cross-posted on Data Innovation Day
Tim Callan is the Chief Marketing Officer of RetailNext, one of the leading provider of in-store retail analytics. I asked Tim to talk with me about how companies like RetailNext are bringing the type of data analytics traditionally used by online retailers to brick-and-mortar stores.
Castro: How are retailers using in-store analytics today?
Callan: Think of in-store analytics as the equivalent on online analytics offerings like Google Analytics or Omniture, but for brick-and-mortar stores. Some of the common uses of in-store analytics are floor layout optimization, staffing optimization, theft reduction, measurement of marketing and merchandising programs, testing display and fixture effectiveness, optimizing checkout queues, and monitoring for stock-out situations. The bottom line improvements have been quite dramatic. For example, Montblanc and American Apparel have reported that they used the RetailNext platform to improve same-store sales 20% and more than 30%, respectively. Brookstone used the platform to reduce shrinkage by about a million dollars a year. And Family Dollar remodeled more than 1300 stores in nine months based on the insights it learned from RetailNext.
Castro: What kind of data is used for in-store analytics and how is it collected?
Callan: Philosophically we believe that all data sources are potentially useful, and we strive to gather and make sense of as many types of data as we can. The most common sources are video streams from in-store cameras and shopping basket data from the store’s Point-of-Sale (POS) system. That’s because these two sources are so incredibly useful. In-store cameras can provide information about how customers move around the store – not just how many shoppers visit the store but where inside they walk, where they stop, and where they look. And of course the POS system is the source for actual sales that take place. Combining these two sources provides the ability to calculate a huge number of KPIs such as overall shopper conversion, traffic by area of store, traffic to “dwell” ratio by product (what percentage of store visitors stop adjacent to a specific product), dwell conversion (what percentage of these people purchase something) and many others.
That said, we can add more insight by taking advantage of further, additional information sources. For instance, integration with the time-and-attendance system can show retailers how to optimize staffing for maximum conversion. Or tying into smart shelf sensors can add further detail to store layout optimization. RetailNext can incorporate information from promotional calendars, RFID, staffing software, Wi-Fi-based employee trackers, and even the weather.
All this information is then available in real time in whatever form the user specifies. Managers can access their data through a flexible web-based dashboard, configurable reports that come straight to their inbox, instant alerts for predefined conditions, a mobile application, database-compatible exports, or third party systems.
Castro: This type of technology seems to be relatively new. Why is this possible now?
Callan: While the e-commerce world has extensively measured store performance for more than a decade, it’s only in recent years that computer science has reached the point where the same is possible in physical stores. This breakthrough has been enabled by the ability to process large quantities of unstructured data in real time. In-store analytics solutions like RetailNext are a good example of “Applied Big Data” – which is the use of massive data storage and processing technology in purpose-built systems that address common problems previously unavailable to technology solutions.
The physical world in particular has been rich ground for applications of this type, with focus on major components of our society such as traffic, power grids, commercial airlines, and now retail stores.
Castro: How do these types of in-store analytics benefit consumers?
Callan: It’s great for the consumer because store operations now have more information to create an environment that matches their customers’ demonstrated desires. It’s one thing to walk around stores and make chance observations or to gather the opinions of a small number of customers that fill out surveys or otherwise provide feedback. These methods are fraught with bias, error, and inefficiency. Instead, retailers can now gain first-hand, statistically significant knowledge of how their entire shopper bases behave. They can learn from these observations and even perform tests, all of which allows them to make the changes the entire set of shoppers prefers to see. And because the solution tracks shopper behavior on a store-by-store basis, retailers can optimize their responses for different regions, store types, community demographics, or any other factor they can identify.
Castro: What’s next for the future of in-store analytics?
Callan: We’ve just scratched the surface of what’s possible. We believe in the long run that the actionable insight available to brick-and-mortar retailers will far surpass what is available to those in the online space. We continue to add new input sources and enhance what’s possible with the sources already in place. For example, we’ll soon be rolling out highly reliable demographic classification of store visitors into males and females, a very useful data point for retailers seeking to better match their offerings to store visitors. Over time retailers have increased the number of systems in their stores that constitute potential data sources, and retail megatrends like mobile POS, in-store Wi-Fi, mobile applications, and “omnichannel “ (combining all channels into a single, integrated customer experience) all will require analytics for retailers to understand and optimize them.
“5 Q’s on Data Innovation” is part of an ongoing series of interviews for Data Innovation Day by ITIF Senior Analyst Daniel Castro. If you have a suggestion for someone who should be featured, send an email to Daniel Castro at [email protected].
