Will Web Analytics Work For Mobile? Home Depot IT Chief: Most Retailers Are Behind, Think "We'll Get To That Later"

As retailers move to embrace mobile commerce, there are debates about what types of analytics should be used for mobile and even whether mobile analytics—or any single-channel effort—is necessarily a good thing. Most retail IT leaders, including Home Depot's Senior VP/IT and one of her counterparts at HSN, say that many chains are so early in their mobile thinking today that such debates are premature.

"I think mobile is so young that we're not sure yet. Our analysis is developing in that area," said Home Depot's Cara Kinzey. "And I think that retailers are behind, [with many saying] 'We're more concerned about sales and we'll get to that later.' Honestly."

Sean Bunner, HSN's Operating VP, echoed Kinzey's sentiment. "It's such an early channel to get so granular. There's some overall trend stuff we're more interested in, like 'what category of merchandise are they purchasing?' From what we've seen, mobile is significantly different than Web or, for us, TV," he said. "But even within mobile, between mobile Web and apps. You see pretty significant shifts in categories, so we then have to get into CRM activity to see 'Why is that?' Do you want to merchandise that store differently?"

Bunner's and Kinzey's comments were part of a StorefrontBacktalk IT leader panel that examined the various issues facing retailers looking to aggressively integrate mobile. Much of that panel was used for a series of podcasts, including "Another interesting point is your marketing resources. All of the print you may still do or E-mail, the touchpoints you have with your customers. Mobile allows you to have a call to action," he said. "I'm surprised I don't see more of that. 'Text in for something.' It's a marketer's dream to know. Now you know they are holding this piece of literature you sent them and they responded, and then tying that back into your CRM."

Analytics, despite its name, is rarely truly analytical. It is more akin to Data Collectors, with the "analysis" they do actually more like intelligent sequencing and flagging patterns and activity that defies historical patterns. The real analysis still has to be performed by the retailer, whether it's someone in IT, marketing or a line-of-business executive. It's the analysis that tells senior management what these numbers mean and what they should do about them. And with mobile, that can get tricky.

For example, many sites want their customers to spend as much time on their site as possible, looking at and considering their merchandise. In general, that's a desktop function. On a mobile device, it's more of a "tell me what I want to know" or "sell me what I need now" scenario. As Pizza Hut CIO Baron Concors argued, on mobile, more is quite often not better. "When I see long times on our pages, it's not necessarily a good thing. Sometimes that can mean you have people who are really confused about how your Web site works and you have to go much deeper, on a page-by-page basis, and see what people are clicking on, what pages they are going back to, going forward to, who they're leaving your site for, where they came to your site from. All of these things come together to tell a story that you've got to get deeper on," Concors said. "On the mobile side, I think it is important to understand conversion rates, traffic, where people are dropping off the application and things like that. We're probably a little bit ahead in that game, in the sense that a lot of the vendors that do Web analytics are just now offering this for mobile devices, so I think you'll see a greater focus on it going forward."

Another panelist, Ann Taylor CIO Mike Sajor (he was just promoted to CIO from CTO), took a very different position. He challenged whether mobile-only analytics is a good thing, insofar as it reinforces the disjointed multi-channel—as opposed to the integrated merged channel—philosophy.

"I'm a little uncomfortable with some of the things I'm seeing emerging around the differentiation of channels in analytics. To me, there's an opportunity for cross-correlation and doing intelligent analytics across the entire behavior of an individual. Last I looked, I am one person. I represent myself through different personas in different channels at different times. Why would it make sense to overanalyze within the constraints of a single channel what my behaviors are?" Sajor asked. "What I am looking for are analytic engines that help me understand behaviors across these different channels but don't try and pigeonhole me in one of them, because that's not how I appear."

Not only is paying special attention to one channel not necessarily helpful, it can actually be misleading. Sajor detailed how he envisioned an end-to-end customer picture—the much ballyhooed single view of the customer—working in a mobile reality.

"Data behaves in funny ways, and it tends to crop up where you least expect it. If a client in a mobile context does not convert, she leaves that shopping cart behind. The reason why that conversion didn't happen could crop up somewhere else. So it's important to have visibility into that complete stream of data," Sajor said. "For example, that reason could crop up in a Facebook comment about how she happened to be near the store, walked in, looked and said, 'Gosh, that's really a terrible pink and I hated it. And it didn't look anything like what I saw on the mobile device or what I saw on the Web site.' So it's a question of looking across the complete compendium of sources and understanding what they are and where she is likely to turn up. Looking at that complete data stream and doing that correlation. That's probably a step beyond us today, but I think it's pretty close."