Wal–Mart Study Shows 13 Percent Inventory Accuracy Boost With RFID
Written by Evan SchumanMarch 14th, 2008
This is one of these “everyone already pretty much knew this but it’s nice someone took the time to prove it” studies. The University of Arkansas was paid by Wal–Mart to find out whether RFID actually improves inventory accuracy. It found that, yes, it does, to the tune of some 13 percent.
But the methodology is a lot more interesting than the as-expected result. The team followed air fresheners at 16 Wal-Marts for 23 weeks where “a national inventory-auditing group determined each day’s on-hand air-freshener inventory by manually counting every item in all 16 stores.” Read this RFID Journal piece for a look into how to prove inventory accuracy and why a retailer can never afford to duplicate this study.
-Marc
March 14th, 2008 at 10:58 am
It’s deja vu all over again. Let me be the first to predict that when the final report is released, a new analysis of the data will reveal that the 13% boost in accuracy really was more than 26%.
March 16th, 2008 at 8:06 am
I had time to browse through the report and found the following:
“Combined with point-of-sale data, a much more accurate view of inventory — both on the shelf and in the backroom — can be provided. As product is sold, PI can be updated based upon a knowledge of items on the shelf (from point-of-sale data) and RFID-generated information of product in the backroom (i.e., tag reads in the backroom).â€
Later in the report the following appears:
“Test stores were provided with a ‘PI-adjustment’ (i.e., auto-PI) system that automatically adjusted understated PI.â€â€¦. “Control stores were left untouched (i.e., no
auto-PI adjustments) “
It appears that the main difference is the change to current practice for the test stores only. That is; the use of POS data to drive the replenishment cycle. This was also the case in the previous study 3 years ago on Out of Stocks (OOS). If this change is removed, there is no difference between RFID stores and non-RFID stores.
The big question is “Why Wal-Mart did not adopt the use of POS data to drive the replenishment cycle after the first study?â€
I doubt that this increase in accuracy has anything to do with RFID. This study is in desperate need of peer review. I suspect that if the non-RFID control stores used POS data to drive replenishment their results would be the same.
March 19th, 2008 at 12:25 pm
Having clear ROI will certainly help drive adoption, but in newer markets, that’s sometimes not enough.
Not all companies are as technically progressive as Wal-Mart. Many are “later” adopters, or budget and organizational issues get in the way of progress.
For broader industry adoption, the FOCUS of the ROI needs to be aimed at issues that generate a high degree of pain for the company. In fact, when there’s a highly painful business problem, companies often deal with it ahead of an official “ROI” study. It’s amazing how the right pain can garner alignment and support quickly. ;)
Sam Liu
Intelleflex, Corp.
March 19th, 2008 at 12:59 pm
Retailing is a tough business that makes for a very challenging testbed for RFID. Low prices, high product velocity, non-technical associates, and thin product margins cry out for very inexpensive and simple systems. Ideally, the systems would be cheap enough to install throughout the retail areas of the store as well, so that shelf stock levels could be directly measured rather than “inferred” from POS, stockroom, and box baler signals. While more powerful (and complete) systems are being deployed in other industries with decidedly more favorable economics—for example in pharmaceuticals, where a single life-saving item can cost $1,000—we’ll know that RFID has truly reached commercial maturity when it becomes an expected part of retail operations.