Google Trial Sends Home Depot Shoppers Away To Lowe's

A mobile vendor who was testing out the in-store Google Maps application this week at a Home Depot store in Florida discovered an unexpected result. While standing inside a Home Depot—which is one of a handful of Google partners on this project—and just feet away from the store's paint aisle, the tester called up the store's inside layout and asked the app where the paint aisle was. The Home Depot partner app quickly responded: At the Lowe's store three blocks away.

It turns out the search field within the Google Maps page is not designed to search the maps but to search Google Local. And even though the app knew that the customer was inside a Home Depot and that Home Depot sold paint (the word "paint" is clearly visible on the Google Maps Home Depot screen, right next to the arrow representing where the consumer is standing), it sent the customer away to a direct rival—even being so courteous as to display a street view that literally shows the customer leaving the Electrical Lighting area of Home Depot and driving down the road to Lowe's. In short, even the app that created the app understood the consumer was standing inside a Home Depot at the time.

The idea of having Google helping a store navigate consumers around that store is a good one. But it's becoming clear that retailers need to be thinking about—and asking—a lot more questions.

The first problem is making sure everyone understands what the app does and what it doesn't do. It can support a small number of categories, but it doesn't currently support any type of text search. For some stores with smaller footprints, having the user simply cursor-navigate around the map is a fine and efficient way to search for a desired aisle. But for a larger footprint store—think of Costco or Sam's Club or even a large Wal-Mart or Target—finger navigation is too inefficient. Why not enable someone to type in the name of an aisle and be directed to that part of the store?

Well, for starters, programming. The retailer-generated floor map would need to metatag every aisle with its name, plus as many synonyms for that name as possible, including versions with common typos. That has to happen for every store and for every retailer. Certainly do-able, but not in a day.

The additional text-searchable granularity is also crucial for the next step, where the maps will be able to zero-in on precise locations of products within an aisle. That's an order of magnitude more complex, and Google's current "several meter" accuracy for in-store maps makes it impossible.

But getting back to what exists today, it's not clear why the current Google Local couldn't factor in that this person was asking for paint while standing inside a Home Depot, which sells paint.

Part of the problem here is that this search would be subject to the same lengthy list of Google search parameters that have been driving SEO professionals mad for years. What metatag terms were prioritized on the Home Depot site?

When the same test was repeated inside a local Home Depot in Bellevue, Wash., the "paint" search sent shoppers to Northwest Painting Services, a local merchant. We do note that Northwest seems to be displayed very high on the page.

The company that performed these tests, PointInside, reran the tests and asked for flooring. Yet again, Home Depot customers standing inside a Home Depot were sent to other stores, even though Home Depot sells the desired products.

Home Depot wouldn't comment for this story, although we imagine its executives are right about now commenting quite a bit to Google. Google also would not comment."Once the retailer has done the critical work of getting a customer into their store, the retailer needs to know that their ads and products will be shown," said Todd Sherman, PointInside's chief marketing officer. "A customer inside a store is a high-value target for advertisers and a lucrative revenue opportunity for Google. It is likely that Google's indoor maps will operate with Google's very successful AdWords model, where algorithms decide which ads to present and are biased toward the highest bidder—which can easily be a competitor."

Note: PointInside sells its own search-inside retail product. We examined extensive screen captures and videos of its results and spoke with others involved, and we're comfortable—for now—with the validity of its experiments.

One CIO of a large national retail chain that is aggressively involved in Mobile Commerce, who asked that both his name and his chain not be identified, said that he sees this as more of an oversight on Google's part than anything evil. "Google usually isn't that evil. That's a bridge too far, even for them. This speaks to glitch more than intent. It's not being indexed."

The CIO argued, though, that these incidents are good in that they force retailers—at a very early stage—to think through the implications of things like mobile in-store mapping. "You're trying to map a bunch of really complex stuff together all at once. Are the rules of engagement even specified? You're providing a map of the store. What's the responsibility of the search provider? The first thing I'd want to do is understand what the contract says."

The fundamental problem here, by the way, is not actually with Google Maps, Google Local or how Home Depot coded the maps it delivered to Google. It's with the lack of available information about local inventory. eBay is struggling with this now, and it is only the latest to deal with the nightmare that is local inventory.

Has Home Depot, for example, taken all of the products it codes for its site and made them all available to the Googles, Yahoos and Bings of the world? Until it has, searches that legitimately should should be sending customers to Home Depot (or, in this case, letting them stay there) will go to rivals whose product codes are more easily search-engine accessible.

That doesn't explain away all of the strangeness from this Home Depot Google trial. But until it's solved, the real problem will remain.