Paul Boyle, CEO of Retail Insight.
If you peek behind the curtain of grocery retail, there is a deflating inevitability affecting every balance sheet: retailers don’t have the product availability they think they have. Otherwise known as Phantom Inventory, this imbalance needs a data-led solution that enables targeted action to get the full potential from every store, says Paul Boyle, CEO of Retail Insight.
The reality is that on average, inventory records are only 50-60% accurate. This might vary depending on the grocer, but the end result is the same: lost sales and decreased customer loyalty. There is also the additional consideration of the cost of labour hours spent on manual, undirected corrections.
This inaccuracy is known as Phantom Inventory – or stock which appears to exist, but doesn’t. By its nature it is inevitable within the retail environment as it’s caused by theft, improperly scanned items or short deliveries. An analysis of the experience of major grocers in both the U.S. and the UK has highlighted that Phantom Inventory can cause as much as 80% of out of stocks (OOS) on the store shelf.
Retailers that base their decisions on inaccurate inventory records see a huge knock-on effect on many of their key operational performance metrics, such as on-shelf availability, sales, waste, shrink as well as capital and labour efficiency.
What they need is greater visibility into the estate so that instances of Phantom Inventory can be surfaced. This is achieved by identifying where there is assumed stock but where there are no sales over a period of time. Store associates would then be alerted to items suspected to be a victim of Phantom Inventory, to take immediate remedial action. This insight therefore ensures that colleagues can reduce and control Phantom Inventory easily, quickly and without manual audits. This is important because manual audits are out of date as soon as they are carried out, and become particularly problematic when audits are undertaken infrequently.
Through a consistently accurate inventory record, retailers can highlight potential future inventory problems prior to these products running out. By combining current sales volume, lost sales volume and on-hand quantities, in trial stores, retailers were able to achieve a 75%+ actionability accuracy. At the same time, they could identify other issues with merchandising and item set-up.
As well as sales gains, lower out-of-stocks and greater customer satisfaction, there are also labour savings in this approach because staff are acting on directed alerts, rather than spending time searching and investigating lines.
For example, a major UK supermarket chain decided to try and fix the problem with lines that should have been selling better because their current inventory record showed good stock levels. Their focus was intensified by the fact that these items had a high lost sales volume attached to them. To help solve the problem, Retail Insight applied its InventoryInsight solution, which took the retailer’s existing data to identify instances of Phantom Inventory.
InventoryInsight uses cognitive technology, based on proprietary machine learning and human-led, advanced analytics, to deliver better execution. Cognitive technology removes the gap between data and execution to enable humans and tech to work more collaboratively. And by delivering accurate and actionable insights in near real-time to every part of the organisation, the traditional silos of data and analytics that can hold back execution are removed.
By adopting this approach, the retailer was able to adjust for Phantom Inventory, alert store staff daily and relate these to store type – convenience, supermarket, superstore. Our alerts consider accurate perpetual inventory, replenishment prompts for stock not visible, stock on promotion and display issues. In pilots, which included initiating a more collaborative interaction with suppliers, increasing safety stock of certain high-margin lines led to a 25-50%% jump in sales.
Across multiple implementations in both the UK and U.S., strategies that remove Phantom Inventory will take availability to at least 80%. One large U.S. grocer, for instance, saw a +102bps increase in availability while another reduced out of stocks by 40%.
When it comes to cognitive technology solutions like these, time to value can be weeks, not months; however, it is always advised to proceed at a pace that allows teams to feel empowered. This approach provides feedback that gives the early results much higher visibility across the organisation and provides encouragement to other teams.
Leveraging data-led insights will make major inroads into the problem of Phantom Inventory and do away entirely with the need for inefficient manual audits. Supported by cognitive technology, teams will be able to collaborate more effectively and get rapid speed to value using partner domain experience. Projects undertaken for retailers in Europe and the U.S. have delivered significant returns within months, and embedded technology now provides a platform for long-term, sustainable benefits.
