Walmart’s Inventory Struggles Offer Valuable Supply Chain Lesson

Renowned for its outstanding supply chain, Walmart has suffered a string of bad publicity. Reports of excess inventory coupled with empty shelves have tainted the retail giant's reputation as a flawless executor of supply chain management. Walmart's inventory issues remind us past achievements are no guarantee for future success.

And based on a visit to my local Walmart, where empty shelves greeted customers in the middle of the holiday shopping rush, proves there's still a lot of work to be done before the blue-clad Walmart army will again “march in sync” as the new Walmart U.S. CEO Greg Foran put it.

That a company which has pieced together an information technology infrastructure larger than that of any other private company in the world still faces inventory issues provides a stark reminder of the fickleness of success. Resting on past achievements is not an option even when you are the top global retailer.

Keep in mind Walmart's logistical and operational triumphs have to a large extent been built on the successful implementation of technology to manage its mammoth operation of $32 billion in inventory and 11,000 stores worldwide. It was the unprecedented collaboration between manufacturers and Walmart's suppliers that made the company, despite its size, work like a single firm.

The retail giant early realized the importance of sharing information across the supply chain, creating a central database, Retail Link, which displays information as detailed as real-time sales data from Walmart cash registers. A collaborative planning, forecasting, and replenishment scheme enabled manufacturers and suppliers to synchronize their demand projections while staying connected through technology such as store-level point-of-sale systems.

Faced with the recent inventory debacle, Walmart once again seems determined to seek a large part of the answer in technology, rolling out a new system for replenishment and inventory control. The Global Replenishment System (GRS), a downward forecasting model, is the polar opposite of the forecasting engine that it is set to replace, Inforem, a bottom-up system developed by IBM. Designed to provide an improved real-time view of inventory, GRS also places added emphasis on promotional/event planning.

Foran said recently simplifying inventory management process actions will further address the fact that inventory has been growing at twice the rate of sales, causing SKU overload at local stores. There is no quick fix since the issue at hand is so complex.

An interesting observation in Forbes by retail analyst Paula Rosenblum identifies as the root problem “the tyranny of turn” or the “velocity of sales as a percentage of average inventory.” Rosenblum argues the solution could be found in new RFID readers for wide-area monitoring combined with a few handheld devices to circumvent metal interference. The addition of payroll – Foran himself admits the company has “cut muscle instead of fat” – would also contribute to get the backrooms back in order.

Despite the negative news reports, Walmart is still one of the world's most efficient retailers. What do you think the company needs to do tackle the inventory issues?

4 comments on “Walmart’s Inventory Struggles Offer Valuable Supply Chain Lesson

  1. Brian Fellow
    January 26, 2015

    Promotions are a significantly more prominent part of the retail landscape than they were pre 2000. The growth in promotional spend has been catalysed by an increasing plethora of Digital marketing techniques and tools that enable companies to almost instantly reach the end consumer. The result – a legacy bottom up IBM forecasting methodology is a poor fit to manage inventory in the modern retail world. However, let's not throw the baby out with the bathwater. Some of the “IBM” models will be excellent – they just don't adequately deal with promotions and events.

    If you cannot accurately predict the impact a promotion has on inventory, then service levels will drop and inventory will rise. Products such as Oracle's Demantra, and SAPs equivalent in the same space, do not have algorithms that match promotion cycles accurately enough to demand. They are not dynamic enough. The devil is definitely in the detail, and the promotion demand curve fitting predictive algorithms do matter. 

    Take a simple inverted S curve for example, for a price only promotion. Price on the Y axis, volume on the X. Up to a point (let's call it the 'point of inflection'), reducing price results in INCREASING volume at an INCREASING rate. However, as you continue to decrease price beyond this “point of inflection”, there becomes a saturation point where people purchase more product but at a DECREASING rate. The market is saturated, and the law of diminishing price elasticity applies, if you like. 

    If you can't model these promotional curves accurately, and all their characteristics including cannibalization of other brands as you promote, impact on promotion clashes, impact on stock buy up for a promotion and the impact on a reduction in baseline demand following a promotion as consumers have gorged themselves, then you are going to spend increasing amounts nowadays maintaining service levels as a retailer. The right management science behind promotion and event management is foundation knowledge for a wholesaler and retailer in an environment where promotions are critical to drive online and bricks and mortar foot traffic using digital marketing techniques. Outsourcing this intellectual property to an eager software vendor may result in you being tied into solutions that were fit for purpose 10 to 15 years ago.

    To me, many retailers are literally scared of the maths and the modelling – view it as academic pontification when “a few good people and common sense” will prevail. No – what prevails is being able to accurately model the impact of promotions and deploy inventory in advance or “NRT” (near real time). The techniques are there – but you need to be willing to apply them. 

    Data Cleaning is another key consideration. You have to be able to quickly identify outliers, dirty data, clean it up, and proceed. Basically, you have to know your business at a much more intimate level than what was the case 15 years ago in terms of demand management

    It is OK to talk about woolly concepts such as “demand and forecasting learning algorithms”, but to put them in place in a data rich dynamic environment is another thing altogether. There are (a few) leading wholesalers in ANZ that I know who will be more advanced in this area than Walmart, having read this article. They are only more advanced because the retailers have demanded more and more investment in trade promotions due to the reasons outlined in the first paragraph. There has been an increasing recognition of the need to measure and improve on that return on trade promotion investment. In ANZ, the driver for investing in the required management science and technology has not been inventory, it has been the amount giant Oligarch retailers have been demanding to fund increasingly aggressive promotional campaigns.

    Let's hope Walmart has invested where they need to here, as easy for a savvy software vendor to pull the wool over one's eyes by stating “we have been around for 10 years – we are proven”. The fact the core algorithms have not been changed for 10 years is a cause for great concern, not a comfort! I, for one, was disappointed at the level of service Walmart provided this Christmas. Let's hope they improve moving forward.

  2. Hailey Lynne McKeefry
    January 28, 2015

    @Brian Fellow, thanks for weighing in. You make some important points. More data, less time… it's going to be a challenge for the industry.

  3. Brian Fellow
    January 28, 2015

    Yes – more data, less time. The bridge is technology.

    If the investment in the correct use of technology to process the additional data lags the change in operating conditions that caused the increase in data, which in this case corresponds to rich information relating to promotion effectiveness, then you will end up with more data, less time and less service.

    The opportunity (upside) here for Walmart is huge, but investment in realization of that opportunity unfortunately has lagged structural changes in retail such as the rapid growth in :

    1.     Digital as a means to reach consumers 'near real time'

    2.     In store promotion and event activity

    These structural changes are a global phenomena that most wholesalers and retailers face. Walmart simply faces these challenges on a grand scale, hence the impacts of not adequately focussing the required investment in promotion effectiveness learning is more acutely felt.


  4. Arnaud Kleinveld
    February 2, 2015

    Great reply Brian. From our experience we know that in most cases it's a lack of knowledge. Instead of being scared for maths and modelling our clients are simply not aware of the insights that technology of today can provide. And of course we see it as our task to keep our clients up to date with the latest (forecasting) techniques. However, it's often a daunting task to convince our clients to allocate enough resources to every single project. The impact on the supply chain caused by digital marketing techniques is seriously underestimated. My opinion is that Walmart should invest more in the knowledge of its people and in sourcing for the right solution.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.