In-Store Trial: 3 Mobile Datapoints To Locate Customers
In a five-store trial—slated to expand chain-wide in the next two weeks—the Meijer grocery chain has gotten creative about letting customers locate products on the shelves using their phones. Given that GPS won't work in-store and that in-store hardware sensors are expensive and labor-intensive, the chain is using a combination of Wi-Fi signal strength and product-barcode scanning to zero in on the customer's location.
The potential of this microlocation mobile approach is compelling, because it provides a relatively easy—and somewhat accurate—way to help customers find product. Of course, that's not the goal of all chains. Some chains—such as Costco—depend heavily on the customer stumbling on impulse buys as he/she wanders the aisles in search of the elusive clothes pins or peanuts.
The Meijer FindIt App trial—which is using technology from a vendor called PointInside—tries to deliver mobile product-discovery using several simultaneous data sources. On their own, the accuracy is fairly weak. But when combined, the system gets a fairly precise sense of the customer's location.
The first datapoint is the store's planogram, which the chain will need to regularly update. Customers can see it on their phones and, theoretically, can look at an aisle heading and figure out where they are on the map and then navigate from there. This is the low-tech "I am here" concept from shopping mall maps.
From there, the system accesses the store's Wi-Fi network. Critically, though, customers don't need to find the network on their phones and also don't need to try and log into the network. The sole purpose is to proximate—within about 10 meters—the customer's location based on Wi-Fi signal strength, said Todd Sherman, PointInside's chief marketing officer. "We'll triangulate. That's the Wi-Fi fingerprinting that we do," he said.
Still, a plus—or minus—of 30 feet in a grocery store can make finding that bottle of soy sauce with your iPhone not that helpful. The next datapoint is a product scan. Once the customer can be persuaded to scan some product on the shelf, that product's barcode will—courtesy of the planogram—reveal a much more precise indication of the customer's location.
If the customer happens to scan an errantly placed product—such as a container of ice cream sitting in the paper goods aisle—the system will see the disconnect between the Wi-Fi data and the barcode and then ignore that barcode and await the next barcode, Sherman said. "We can tell when something is out of place," he said.
The third datapoint comes from the phone itself.The third datapoint comes from the phone itself. Once the system has a fix on the customer's location, the system accesses the phone's compass and gyroscope to track the customer's movements, continually updating the current location. Some movements provide another strong clue about precise location, such as turning around a corner. There are only so many places in the store where someone could legitimately make such a move. That data layered on top of the other datapoints can provide an increasingly precise customer location, to be mapped against the hopefully accurate planogram.
As gyroscopes on newer phones improve, the accuracy of these methods will likely also improve, Sherman said, adding that he's hoping to see three-meter accuracy in the very near future, mostly as sales of current-generation smartphones grow.
Meijer, which has 197 stores in Michigan, Ohio, Indiana, Illinois and Kentucky, did not respond to a request for comment by Wednesday (Oct. 26).
The Meijer app initially will permit someone to search for a specific item. But future versions will enable a customer to post a full shopping list, where the app can create the most efficient store path to obtain all of those items. Like most shopping list apps, the potential for suggested items will then be almost irresistible. ("I see you like overpriced items. May I show you some French sea salts?").
But that's for the future. Right now, the app's primary approach is to take a specifically-sought item and find it on the map and then direct the customer to it. “If you’re looking for peanut butter, it will drop a pin on the map, within a couple of feet,” Sherman said.
The problem is that customers will not likely use it just to find generic peanut butter. They’ll want to use it to find Skippy Peanut Butter, Extra Chunky, With Low Sodium. When I am dispatched on a shopping trip and have an unfamiliar item on the list, I can generally find the aisle quickly and even the rough area, so that it will get to me to “Peanut Butters” or “Dry Cereals” or “Shampoos.” The value of a mobile app is at the next level of granularity.
Instead of staring up and down through six shelves of similarly-looking product hoping to find an exact match with my shopping list, the valuable app would zero in on the item and focus a laser beam light on the exact box.
And if that exact product isn’t in stock, it would save me the time of looking for it and say, “We’re out of that item and are expecting another shipment tomorrow morning. Would you like to reserve a jar? Would you like me to point to a very similar brand? Would you like me to write a note to your wife certifying that you really tried to find it?”
That’s where the distance between a few meters, a few feet and a few inches makes a huge difference.