or Search by TOPIC
Search Supply Chain Videocasts
  Sign-Up Free Newsletter
First Thoughts
  By Dan Gilmore - Editor-in-Chief  
  September 10, 2009  

Supply Chain News: Warehouse Management to the Rescue for Retail Out-of-Stocks?



This may be one of the oddest columns I have ever written, but I think Warehouse Management Systems (WMS), or at least one of the core concepts, could solve some problems in retail stores and – believe it or not – in our healthcare system.


As some of you know, I have quite a bit of experience in the WMS area, and must confess to having a soft spot in my heart for this type of software – as challenging as WMS implementations often can be.


Central to the operation of a modern WMS is the concept of “task management,” in which the WMS understands the work that needs to be performed, what operators are capable of doing it, and where they are. It then doles out the work electronically (radio frequency terminal, voice system) to the best worker for the task, and receives confirmation that the work was performed. Then, it’s on to the next task. This is happening across dozens or even hundreds of workers simultaneously in a building.


Those next tasks can also be “interleaved,” meaning that different types of tasks can be combined to optimize efficiency. Do another task next to one you just finished, for example, rather than travelling a distance to repeat the same type of task.


Gilmore Says:


A task management engine would simply bring order to the chaos.

Click Here to See
Reader Feedback

So, I am thinking about the retail “out of stock” problem, and how some huge percent of the time, the goods are actually in the back room when the customer finds no product on the shelf.


The industry, in large part, has lately viewed RFID as the answer to this particular problem, and Wal-Mart made a run at it with its program. There is both the general issue of shelf-level stock outs and the frequent failure to execute promotional displays effectively (on-time, where they should be, etc.). Earlier this year, Procter & Gamble famously ended its pilot with Wal-Mart on tagging displays after calling it a clear success – with the implication that Wal-Mart was just not acting on the data.


There is, in fact, a growing form of “task management” in retail, but it is at a much higher level of granularity, but even that apparently is paying big dividends, according to reports. There, it is tied to employee scheduling to optimize what staff is needed where and when, based on sales history and other factors. I believe it can also direct staff at a high level to do some tasks – say receiving in the back room, or cycle counting. But not at the level of real individual tasks – more like as “blocks” of work.


Last time I looked, most decent sized retailers had lots of RF terminals lying around in their stores – and I mean literally lying around. They simply do not appear to be used very effectively or consistently, from what I see. That’s a management issue.


Could we not largely solve the out-of-stock problem by leveraging these under-utilized devices and adding a type of task management system, maybe morphed from a WMS, to direct workers to get product on the shelves?


What it would take is:

  • A perpetual inventory (PI) system at the store level (largely in place);
  • A basic inventory location system in the back room (sometimes in place);
  • Wireless terminals (largely in place); and
  • The task management engine, integrated with the other pieces (or maybe all from one source).

Retailer Target said early on in the RFID saga that it has a light form of this – its system generates a type of “pick list” for shelf replenishment, based on POS. But that is a paper-based system, and not really capable of dealing at a specific task level (or so I think – haven’t really seen it).


Pretty simple – the PI says a location needs replenished (at whatever level is set for that SKU), and a specific task is set for the move. It goes into the queue, and is prioritized based on other variables (just for example, perhaps high-margin items generally go to the top of the queue); if more units of that SKU are sold, meaning it is increasingly in danger of going to zero at the shelf, the task rises in the queue.


Stock personnel don’t have to wonder what to do – their terminal tells them exactly what to do, and they confirm it is done electronically. By the way, they could be asked to do a cycle count when they do the shelf replenishment.  If they see an empty shelf or peg, they could also do a cycle count, and/or check to see if a replenishment is on the way.


We did a story awhile back on how outdoor retailer Cabela’s had implemented a light version of their DC WMS inside their sprawling retail stores. The drivers there were physical store size, the huge amount of SKUs they carried, and the fact that often there were just 1-2 items of a SKU on the floor. Made perfect sense. (See Cabela's Uses WMS to Drive DC Efficiency across Distribution Channels.)


I don’t know if a true “WMS” is really needed, but the basic, highly-proven concept of the WMS task management engine certainly could be. The wireless terminals are largely in place. The bar codes are largely in place. A location management system in the back is often in place, and easy enough to develop if not.


Integrating the PI may take some effort, but I don’t think especially that much.


Early on in the EPC story, Jeff Woods, an analyst at Gartner, saw the best fit for RFID versus bar codes was in “chaotic” environments, without much process discipline. He cited a battlefield or military theater, for example – and RFID has done much to improve material flow there. Interestingly, he also cited a retail back room as another area of often chaotic processes.


Dick Cantwell, then of Procter & Gamble, made much the same point in his famous interview with me about the failure, to that point in time, of any approach to getting displays to the floor on time and correctly.


Is this how it really has to be?


A task management engine would simply bring order to the chaos. It does for material movement in thousands of DCs across the US and the world.


Certainly, a task management engine could work with RFID, and indeed the vision even for the displays was that the RFID data would generate tasks of some sort. What system would manage that was not clear.


But WMS already has a model and technology that could solve the problem quite well, with bar codes or RFID, without re-inventing the wheel.


The WMS would need to be much slimmed down, and I am sure some changes made versus the DC variety to have the right fit.


I am out of space, but, believe it or not, I think the same basic task management concept could improve efficiency and quality in our hospitals. I say that after having visited someone in the hospital this summer, and seeing how uncoordinated and informal the execution of work seems to be. I am not proposing a robotic task management system for everything there, and will admit that I am not clear how the task requirements would be generated, but I have no doubt that someday we will see something like this in healthcare.


Am I nuts on the retail thing? Shouldn’t more retailers be looking at this? Couldn’t we take a big chunk out of the overstock problem at shelf right away, without a whole other industry program that never seems to work?


If any retailers out there are interested in exploring this, I would be happy to help them take a look.


What do you think of the idea of real task management, maybe even light WMS, in retail? Could it solve the out-of-stock problem, at least for those items with inventory in the back room? Have more taken this approach than we know about? Let us know your thoughts at the Feedback button below.

  Send an Email  

Sept. 11, 2009

Dan, you are absolutely spot on here.

I worked for Safeway Stores in the late 1970s and we had developed a tool for store managers to use in planning headcount. Mainframe based, it looked at calendar cycles, store traffic, events, weather and a number of other factors including department volumes to predict the number of employees needed on any given day.

I know these guys think about tasking a lot. Having said that, they (retailers) also obviously know how to manage inventories. This is where CPFR was born, along with the promise of gathering POS data. Things are just things, and things may be in multiple places -- multiple floor locations, in the back room stock, in stales / RTV inventory, etc.

Knowing where the stuff is, is what WMS systems do really, really well. A good POS system, coupled with a good WMS / Space-Stock management system should provide most of what a retailer needs to ensure consumers can find things on the shelf (if it is in the store).

One potential hiccup is that many shelves are stocked by vendors and route drivers. The systems used must be able to collect restock and damaged/stales removal data from these suppliers -- hopefully through common interfaces with their own handhelds.

Steve Murray
Principal and Chief Researcher
Supply Chain Visions


Sept. 11, 2009

Great article as usual  -- a few comments: The solution you offer may work if...

(1) Perpetual inventory was right -- but it is not. A number of CPG companies have proven that in the best cases, 35% of the item counts are accurate. 50% of the items have some form of phantom inventory (think theft and other forms of shrink, mis-picks, mis-checkouts…), and the remaining 15% have hidden inventory.

(2) Products were sold in one single place on the floor -- however, unlike a well-designed warehouse, you may have a promotion display and a home location that are completely different. One of them may be stocked up and the other one is not (there were examples we found where a promotional display actually decreased the total sales lift vs. just operating the promotion from home location as customers were so accustomed to find the goods on the home shelf that it ran out of space and no one actually went to the display).

(3) They had the same perspective: retailers tend to look at their business from a holistic, category-level standpoint and if an item runs out of stock, a 'slide-and-hide' strategy is sometimes corporate policy (a famous club store does NOT have out-of-stocks, should you ask them as they don't care about assortment, just shelf fill rate) whereas a supplier wants to push their own products and view slide-and-hide as the worse practice in the industry

(4) Finally, retailers and manufacturers (and their brokers) were to agree on a distribution of tasks. That may be the most difficult problem.

Altogether, I think that the ability to drive retail execution will end up being a competitive differentiator for CPG companies -- but that it will be primarily driven on a relationship-by-rationship basis, potentially integrating with retailer task management systems when they exist.

Industry leaders are figuring out how to do a better job for THEIR SKUs and avoid brand shifting (or shift towards private labels) by offering better availability using the data coming from their retailer partner.

Cedric Guyot
VP - Marketing
Retail Solutions Inc

Editor's Note:

Cedric, good letter.

I am least concerned about #2 - most warehouses have the same SKU multiple places - easy to handle (piece picking, case picking, reserve storage, etc.). And sometimes those locations are 'dynamic' based on need, and if fact can change frequenty. So I think this could be easily handled within a store.

Other ones are more problematic, I agree, but still have confidence in this. Think the approach might improve PI accuracy, for example.

Dan Gilmore




Sept. 11, 2009

Very good article and I agree with you that a task management system along with RFID and a disciplined work ethic can solve this problem.

K. Rahul


Sept. 11, 2009

I spent a few years working in a hospital when I was in college and observed first hand that, for the most part, hospitals are pretty inefficient.

I have seen little to change my mind in the many years since college. It struck me that there are many redundant tasks and ineffective management of Central Supply (excess inventory, inventory in the wrong location, inefficient purchasing processes, etc.)

With nearly 35 years of Supply Chain Management in industry, I haven’t been able to get beyond a phone interview when applying for Supply Chain jobs in the medical field. I am not saying everything we do in industry would apply, but there is an awful lot of Demand Management, Purchasing, and Inventory Control that would apply in a hospital setting.

So, I guess I am saying that I definitely agree with you on Warehouse Management to the Rescue.

Joseph Kirchner
CPIM Materials Manager
Smith & Nephew


Sept. 10, 2009

Excellent thought.

I totally agree with your assessment. A lot of the problems in retail stores can be solved in this manner. The POS systems in place do not help in this process.They only help in identifying the distribution orders to be raised.

The task of replenishing the locations is definitely a huge task. With the size and scale of the retail stores it becomes a huge task managing that.

I think definitely a lite WMS is a wonderful option.

Ram Ramamurthy Swaminathan
Founder, Director
ACIES Technology Solutions Pvt Ltd


Sept. 10, 2009

There is more activity in this direction than you think.

Manhattan Associates presented at the RILA Logistics conference this year how a SaaS model of their warehouse management systems are being used in grocery and convenience store operations to pull stock from the back room to the sales floor. It is being activity tested, but there are still some roadblocks.

First and foremost is the lack of discipline on a retail sales floor. A superior DC operations has a high level of operational discipline, where there is a clear expectation that the human staff will follow the systematic processes. In the DC culture the associates are trained over and over to follow process, and where the discipline is absent the processes get sloppy.

Now go to the sales floor of any big box retailer and look at the staff in the store. Do you see the discipline needed to make a WMS type of task management system work? Not that it cannot be  -- but the odds are stacked against the process.

It is a great idea -- but a culture remodel that builds process discipline is needed to make the foundation processes work.

David K. Schneider
David K Schneider & Company, LLC


Sept. 10, 2009


Thomas A. Moore
Warehouse Optimization LLC


Sept. 10, 2009

Retail could probably use RFID, but the work we have done says that it is still too expensive to get an acceptable ROI.

I did a very large project with AT&T in their new IPTV program. They have an RFID pilot going on now to track set top boxes which are $100 to $250 items. This RFID program is weighted with technology and cost issues.

So while it has been an interesting experiment and technically successful, the commercial aspects of tracking $100 inventory items is a long way off.

Also, we did an interesting project with a large beer brewer who has a penchant for freshness. The project involved a light WMS to keep track of and rotate stock to help distributors manage FIFO. A simple concept but if you have ever been to a beer distributors warehouse, they have other priorities much higher than running an efficient warehouse operation.

That project did substrate your postulate that in retail (since distributors do act a lot like a retailer from an inventory perspective) there is a place for and value for a WMS. The WMS supports stock management (FIFO), inventory management and in some cases, operational optimization such as slotting (I wonder who decides that the pallet of 100 watt bulbs should be on the top rear shelf at Lowes?)

Ron Johnson

Supply Chain Digest Home | Contact Us | Advertise With Us | Sitemap | Privacy Policy
© 2006-2014 Supply Chain Digest - All Rights Reserved