Target’s Image Recognition App Is Limited To The Point Of AbsurdityWritten by Evan Schuman
Target on Monday (July 21) rolled out an image-recognition mobile app called In a Snap, an effort that simultaneously merits applause for its innovation and boos for its stunningly limited scope.
A true image-recognition app would be trained to look for a huge number of items that would be available from Target and to spot them in the physical world, in any news photo, ad or posted Web image. If a shopper (guest, in Target parlance) is admiring a Reuters news photo of a celebrity wearing a gown that the shopper likes, it could recognize that image and offer up whatever Target has that is closest. Amazon and others have experimented with similar efforts (consider Amazon’s Mechanical Turk project) and Google’s image recognition function within its search engine and Google Glass have also kicked those virtual tires.
Target’s description of In a Snap on its own site implies a similar effort: “Have you ever flipped through a magazine or catalog and fell in love with a product — a rug, chair, lamp, whatever — and wanted to buy or bookmark it immediately? Yeah, us too. That desire for instant access to inspiring product is the idea behind Target’s latest mobile app, In a Snap. Compatible with the iPhone, iPad and iPod, the new image recognition app lets you easily shop Target items directly off the pages of magazines and printed ads — without having to scan confusing codes, look up links or search for product online or in store.”
Sounds good, but the app itself doesn’t come close to delivering. It doesn’t try to recognize images. It’s solely designed to recognize a handful of specific Target ads—ads that are labeled as being compliant with this app. Despite its claim that it can deliver “without having to scan confusing codes,” I’m struggling with seeing how this is any better for its shoppers or for Target. For both code-scanning (even the much-maligned QR codes) and In a Snap, the shopper must download an app, launch that app and point the phone at the ad. The only difference is that a QR code reader can be used for a huge number of applications, whereas the Target app is limited to ads from Target–and a very small percentage of even Target ads.
When the app looks at the Target ad the chain mentioned in its announcement, it doesn’t allow the user to zero in any element of the ad, but merely the ad itself. It then displays short descriptions of 38 products that appeared in the ad and a shopper needs to scroll all of them to try and find the desired image, assuming it’s even there.
“Target is looking to test and learn with this technology. We want to see how guests respond and engage with In a Snap as we consider future uses and integrations,” said Target spokesperson Eddie Baeb. “One of the key benefits with In a Snap is the experience keeps users in the app, even while they enjoy all the options a guest gets on checkout pages of Target.com – such as free store pickup. So the app is built to be simple and make purchasing, sharing or bookmarking easy for guests.”
The problem with Target painting “keep users in the app” as an advantage for shoppers is that, frankly, it isn’t an advantage. From the guest’s perspective, it’s simply a screen that pops up with more information and an easy way—ideally one-click or something as close to that as possible—to purchase the item. Whether that page resides on a Target Web site or a Target mobile app seems irrelevant to the shopper. (We repeatedly asked for specific examples of shopper advantages of this new method, but none were offered.) As long as it’s the same number of clicks or actions, it’s transparent to the shoppers.
This gets into a common issue with apps, which is the tendency to take baby steps, to roll out the program with as little functionality activated as possible, to collect feedback. The problem is that that is ultimately self-defeating. Shoppers will be disappointed—understandably—because the tested app isn’t delivering much value beyond what already exists. It makes perfect sense to limit initial capabilities, but you must provide at least a few concrete improvements. Otherwise, your initial feedback will be both negative and unfair to the app. That said, shoppers can evaluate only what you offer them. If you limit it to the point of absurdity, it’s hard to blame the shoppers.