RFID and Data-Driven Retailing

Jun 18, 2012For last week's article about Impinj's new store simulation system (see Impinj Releases RFID Store Performance Simulator), I interviewed Larry Arnstein, the company's VP of business development. One thing he said struck me: "The larger message here is that retail has to become more data-driven."

Retailers might believe that they already collect a lot of data, but the reality is that scanning bar codes requires labor, which is expensive. Because it costs so much to collect information via bar codes, retail companies typically conduct inventory counts only once or twice annually—and the data they do have is often wrong. For instance, inventory accuracy for apparel stores averages just 65 percent. But the problem goes beyond inventory accuracy.

Retailers lack a high volume of data regarding goods' movements from their warehouse to the back of a store, and then on to shelves. Bar codes are not usually scanned when items arrive at a warehouse, when they are shipped to a store, when they arrive at that store, when they are put away at a particular location, and when they are then picked and moved to the store floor. As such, the retailers often have no idea how long these processes take to complete, or how well the stores are executing these tasks.

Here's something else that retailers don't know: whether an item didn't sell because it was unpopular with shoppers, or because it was not out on the shelf when consumers came in to buy it. This really is one of the most fundamental questions for retail firms, but they are unable to improve merchandising since they can't accurately measure success and failure, or the causes of failure.

Every MBA student learns the old maxim that "You can't manage what you can't measure." And if you can't collect data, or acquire accurate information, you can't measure success. An item-level RFID system can deliver all of the data that retailers require.

I like the fact that Impinj's Store Performance Simulator can help retailers begin to understand the impact not just of RFID technology, but also of adding additional service labor or inventory-management labor, or of shortening their supply chain. Instead of spending tens of thousands of dollars to test results in the real world, they can simply perform a virtual simulation.

I don't know what is under the Impinj simulator's hood, so I can't state for certain how accurate it is, but Arnstein says his company will work with retailers to test assumptions and accurately model their store so that the simulation will be accurate. This means retailers can forecast the impact of using RFID on operating profit, inventory accuracy, unit sales and other metrics. According to Arnstein, this is akin to an aerospace manufacturer simulating stresses on a plane before ever building one.

That's great—but imagine what a retailer might be able to simulate once it had RFID data from live operations. For example, it will be possible, in the future, to collect sales information by location and fixture within a store. This will make it feasible to test how well an item sold at one location versus another at a different store, and to optimize a store's configuration in order to maximize revenue. Retailers do some of this now, but their store configurations are often based on selling price, or on mere guesswork.

It might turn out that jeans sell better on rounders at the front of one store, for instance, and not as well in another. With hard data about an item's location and sale price, the amount of time that product spends on the floor, and other data that an item-level RFID system can supply, retailers will be able to simulate different store configurations and forecast sales. This, in turn, will enable them to make better, more informed decisions.

That's a huge benefit. And having real data regarding shipping accuracy, replenishment times, pick accuracy and so forth will allow companies to measure—and improve—their performance. But that's not the only benefit. The frequency at which RFID data can be collected allows retailers to correct mistakes.

That's one of the things that Arnstein says he learned from using the Store Performance Simulator. "I ran a scenario comparing a store using RFID with a replenishment accuracy of only 80 percent against the same store using RFID with a 90 percent replenishment accuracy," he explains. "The impact on on-shelf availability, front inventory, unit sales and even operating profit is minimal. That surprised me, but it makes sense. If you miss an item today, RFID lets you catch that a day or two later, and to fix the problem."

And that proves that if you can measure it, you can manage it.

Mark Roberti is the founder and editor of RFID Journal. If you would like to comment on this article, click on the link below. To read more of Mark's opinions, visit the RFID Journal Blog, the Editor's Note archive or RFID Connect.