While it is rare to see the topic of inventory inaccuracy featuring highly on the to-do list of many loss prevention leaders, its impact on business profitability can be dramatic, not least as it can lead to items inadvertently going out of stock and therefore not available for customers to buy. As the research on total retail loss identified, lost profits from out of stocks can be profound, especially in the grocery sector. For instance, when data from three European grocers was analyzed, it found that while unknown losses (shrinkage) accounted for on average 0.67% of retail sales, it was estimated that lost profits from out of stocks was likely to be 0.78%. Interestingly, unknown losses were relegated to third place when losses from product wastage were also considered (1.64% of sales).
The impact of stock inaccuracy has also been highlighted by other ECR research which has shown that correcting book stock accuracy more frequently can lead to an increase in retail sales by as much as 6%. For apparel retailers, where SKU complexity is a particular issue (multiple size and color combinations), the issue of out of stocks and their impact upon business profitability has led many to invest in RFID systems to address this issue. Through routine store scanning of RFID tags, out of stocks can be more quickly identified and corrected, especially incidents of ‘phantom’ out of stocks where the product is in a backroom area rather than on open display.
However, for the world of grocery, using RFID at scale is a much more challenging proposition, not least because of limitations on its capability due to product characteristics (use of metal and viscous fluids) and the economics of its use on relatively low value fast moving products. The grocery sector, therefore, largely relies on routine manual counting/observation of stock by store staff to identify and correct inventory inaccuracies—looking for gaps in store displays of products and/or cycle counting. Depending on staffing levels and store commitment, this can be a rather ‘hit and miss’ approach to managing the problem.
Can Video Help?
To help grocers address this issue, developments in video technologies and analytics have begun to offer interesting alternatives, albeit with sometimes mixed results thus far. The concept is relatively straightforward—use strategically placed video cameras to monitor store shelves and use their data to generate alerts when stock levels either drop below a certain level and/or items are no longer present for sale. These alerts can then be used by store staff to prioritize restocking and where necessary, reordering.
While the concept is beguilingly straightforward, like most video analytics, the devil is in the detail, not least the importance of taking into account the complexity of the operating environment and the likely return on investment (ROI). Two issues are of particular note from trials undertaken thus far.
Product identification or location identification? With some grocers stocking in excess of 50-60,000 different SKUs, getting a video analytic to accurately ‘identify’ all these items in a retail store remains a challenging proposition, especially when packaging differences can be very subtle. This has led to an approach which aims to use video analytics to monitor ‘spaces’ on retail shelves rather than the actual products occupying those spaces. This then reduces the analysis to a more manageable binary activity—is there something in the allotted space or not? If a store maintains good planogram discipline (putting the same products consistently in the same space), then this type of analytic can offer useful data to identify when a product is out of stock or not. What it typically cannot do, however, is offer any form of count of the products on the shelf. In this respect, it is acting as an event indicator and not a stock counter.
Location of cameras. Another significant challenge is where to locate the cameras to carry out this form of monitoring. Traditionally, most cameras in retail stores are located up on ceilings or high on walls to avoid tampering and to provide a greater angle of view. However, this can make the viewing of products on shelves challenging, especially when the shelves are high and close together. The alternative is to place the cameras on opposing shelves which enables a better angle of data capture, but does require more cameras and also exposes them to potential interference by customers. It also means that more supporting infrastructure may be required at shelf level (power cables etc.). Both approaches are being trialed at the moment in retailers around the —the former is appealing because it requires fewer cameras and may be able to use existing equipment, while the latter has its advocates because of the quality of the data and the ability to link to electronic shelf-edge labels.
Video Analytics in Retailing: An Evolving Business Case
The use of video analytics to help deal with the problem of inventory inaccuracy is certainly an emerging space although at the moment it seems to be at the stage of offering useful inputs around the incidence of out of stocks rather than providing more detailed data on counts of actual products. But, it is yet another example of how video technologies are beginning to offer additional value to retailing well beyond the traditional security and safety framework that has dominated its utilization thus far. As such, improving inventory accuracy may be another potential benefit to add to a video technology retail investment case.
Video Watch is a monthly column written by Professor Adrian Beck sharing insights on the proactive use and impact of video technologies in retail. It reflects the latest research and monthly discussions of the Video Working Group of ECR Retail Loss, the leading global think tank on retail loss. The research commissioned by ECR Retail Loss is supported by independent research grants provided by Genetec and other leaders in retail loss prevention.