Once the vending machine had evolved beyond the simple "sales stimulator" model, it began to be seen as an extension of -- even a replacement for -- conventional retail outlets. On either side of World War II, retailers experimented with vending as an alternative to human clerks for selling uncomplicated items (like handkerchiefs) that did not require expert advice or fitting. The technology, especially payment, really wasn't up to the task, and the vending industry soon went on to specialize in the provision of refreshments in workplaces and institutions.
Nevertheless, no discussion of "retail automation" can be complete without consideration of vending. There has been a good deal of technological crossover in recent years, from supermarket self-checkout lines to micromarkets, and the latter have opened up a new opportunity for operators. And cashless payment systems that can identify users have spurred the use of vending machines, usually with few modifications, as media for in-store promotions. And a diverse assortment of specialty retailers finds vending an eye-catching means of attracting attention and selling niche products in high-traffic public locations like airports. Both of these developments may catalyze further vending advances.
We were reflecting on all this during a recent visit to the 2013 National Retail Federation exhibition in New York City last month. Our principal purpose was to visit the Intel booth, which not only showcased the latest implementations of the company's Intelligent Vending approach, but also a wide variety of microprocessor-based concepts for retailers of many different kinds. We had little time to explore the show's 500-plus other exhibitors, but half an hour was enough to collect a couple of months' worth of food for thought.
The almost universal availability of affordable wireless networks capable of covering wide areas is being coupled with extraordinary strides in computerized pattern recognition to inspire services that would have been difficult to imagine even a decade ago.
Perhaps the most compelling benefit on offer here is visibility. Intel's "Intelligent Shelf" project is exploring the best ways to capture images of store shelves, updated frequently, and send the pictures to a central computer for processing and interpretation. The results can be used to determine whether the shelf has been stocked with the correct sizes and formats of the products that are supposed to be there.
For the past several years, Vendors Exchange International (Cleveland) has been demonstrating this kind of system designed specifically for vending machines; it uses a camera that travels along the inner face of the vending machine door, scanning each shelf to image the products in vend position. Intel is considering several possible ways to accomplish this kind of imaging in a conventional store, perhaps by mounting little wireless cameras on shopping-carts. We wondered whether the security cameras already supplied with micromarkets might not be slightly modified to provide this kind of image, along with its video surveillance.
In any environment, missed sales can be minimized by the ability to obtain frequently updated -- ultimately, perhaps, real-time -- information on just what is in each slot on each shelf, and whether that is what's supposed to be there. This has been the missing piece in vending machine accountability from the earliest days of multiproduct, multiprice equipment more than half a century ago. At first, the concern was that an item might, through error or faulty instructions, be loaded into a slot designated for a different SKU at a different price.
Alternatively, the product might be loaded correctly, but the price setting for that slot might be incorrect. In either case, the information stored in the management software's database would not align with the results reported from the machine; this would create additional problems if one were analyzing line-item sales data. This has been a seemingly intractable limitation, and it now seems well on its way to being resolved.
We think, too, that micromarket operators could make good use of information about exactly what is being purchased (or picked up, examined and put back in the wrong place) in each location, and of the ability to spot-check the drivers' work in arranging attractive shelf displays.
We also ran across a company, Stoplift Checkout Vision Systems (Cambridge, MA) that produces security cameras and software for supermarket checkout lines, designed (among other things) to detect a clerk who colludes with a customer (or "sweetheart") and slides some items past the scanning sensor so they will not be rung up. It does not take a great deal of imagination to think of analogous uses for this technology in tightening control and minimizing shrinkage.
What seems to us most exciting about all of this is the increase in modularity. It is becoming easier to assemble very sophisticated solutions from readily available components. Coupled with the steady advance in communications technology, this is opening up possibilities that can overstretch the liveliest imagination.