The Extraordinary Challenge of EMS Inventory Management

By Quentin Samelson, Sr. Managing Consultant, Global Business Services at IBM



EMS companies, as readers of EMSNOW know very well. range from some of the largest manufacturing companies in the world down to small, single-facility operations, and are responsible for producing many of the electronic products that consumers use each day, as well as the big pieces of infrastructure that power the data centers and networks that businesses and consumers both rely on. It’s a huge industry – estimated at over half a trillion dollars in 2018, and generally profitable[1] — an impressive statement since EMS companies typically operate in a very demanding environment, and at relatively low margins.

EMS providers face a set of extraordinary challenges in managing their inventory:


  • Constantly shifting customer demand – and a variety of different demand sharing formats and practices used by different customers.
  • Multiple part numbers for the same functional item. Depending on the number of customers they serve, some common electronic parts may be ordered and stocked under dozens of different part numbers.
  •  At least some customers may treat their manufacturing ‘partner’ – the EMS provider – as a no-cost inventory bank, available when needed but with no requirement to accept deliveries just because they had been scheduled earlier. Customers sometimes appear to over-forecast and under-consume as a normal business practice.
  •  A conflict between customers’ lead time commitments (often three months or less) and the lead times demanded by the (customer-chosen) component suppliers (often five months or more).
  • Frequent engineering changes that can result in additional BOM churn.
  • Tense supplier relationships.
  • A combination of legacy processes and systems, often left behind or dictated by customers, with a need for new methods and systems.

And other issues. It can be summarized as a state of perpetual churn, and it can be more than a bit overwhelming, with seemingly every new situation working to prevent effective inventory management. Yet the need for good inventory management is more critical at EMS providers than other companies. Having the right parts on hand to build customers’ products is crucial to an EMS’s long-term success; and not having excessive quantities of those parts is nearly as important! In the short term, surplus inventory creates cash flow problems (and can clutter up stockrooms and manufacturing areas as well). In the longer term, excess and obsolete inventory has to be negotiated with customers – a process that neither party welcomes, and which can often drag on for months.

The good news is that the problem is solvable; the bad news is that it takes work and discipline to keep it solved. But it can get easier over time. The process for EMS providers starts the same as for any manufacturing company: first, Ensure the basics are in place. Make sure you have access to current inventory and demand data, and run an ABC[2] report. It’s crucially important that you set your ABC categories based on value of consumption rather than individual part value or personal preference.

At the same time, you need to set a target inventory range for each ABC classification. For example, you may set a target range of:

  • 1 to 2 weeks for A parts, which represent 80% of expected consumption value
  • 2 to 4 weeks for B parts, which represent 15% of expected consumption value
  • 3 to 6 weeks for C parts, which represent 4% of expected consumption value
  • 4 to 8 weeks for D parts, which represent the last 1% of expected consumption value

Note that you can easily do the math to see what your average inventory would be if you hit the midpoint (or even the max) of each of these ranges. (At the exact midpoint of these ranges, you’d have an average of only 1.89 weeks on hand; even at the max, you’d have 2.52 weeks on hand.)

You can often make a tremendous amount of progress by comparing the actual inventory of A parts to the target range. Encourage individual buyers to resolve major outliers (the “A” part with eight weeks on hand, for instance); this can drive inventory down quickly. Teach buyers to focus on the critical few (the A and B parts) and to manage low-value (C&D) parts by scheduling big, infrequent deliveries so they don’t require much attention. Once many of the parts are in or near the target range, you can work to get each buyer’s average into range, by calculating an overall inventory quality score[3] for each individual. That permits team members to offset higher inventory of some parts, which may come from remote suppliers or have other issues, with lower, more aggressive inventory levels of other parts that may be locally sourced or stocked.

Getting your buyers to “manage to the target range” is an incredibly effective technique, partly because of the discipline it establishes; but there’s another reason that it’s so effective: it exposes the kind of issues and opportunities that need to be addressed at a facility, customer or even a corporate level. For instance: what do you do about the casting, an “A” part, that was tooled at a supplier located on a different continent than your manufacturing facility? It may be too heavy to ship by air freight; but using surface (ocean) freight is likely to result in unreliable deliveries and higher inventory levels. You may need to work through a whole range of options, including tooling an alternate (closer) supplier, evaluating a series of transport routes, etc. Or perhaps you have several manufacturing facilities in the same region, which consume many of the same components. Setting up an inbound logistics hub (often combined with multi-site buyers and supplier-managed inventory at the hub) can benefit all of your facilities, as well as improve predictability for the suppliers – a real win/win scenario.

* * * * * * * * * *

At this point, I’ve had at least some exposure to about a dozen different EMS providers. Certain common issues exist at many of them; each of these can be resolved over time.

1.    Production & Inventory Control Education is often lacking, but it is the foundation of sustained improvements. The body of knowledge provided by ASCM (formerly APICS) is very accessible, and is available in local languages in most countries. Get your people trained.

2.    Basic Execution processes are often flawed or spotty. That’s sometimes because knowledge of inventory control concepts is lacking (see #1). More often, it’s because of lack of performance benchmarks. I’ve visited more than one factory where the production floor was nearly world-class but the receiving area and warehouse were stuck back in the 1970s. More than one had a receiving process that took days instead of hours. So, start measuring performance, finding benchmarks, and improving processes.

3.    Supplier Relationships are often problematic at EMS providers. Component manufacturers naturally want to cultivate a strong relationship with OEMs so they’ll be designed into the next generation of products. An OEM’s demand fluctuations can also contribute to tension between EMS providers and suppliers. But suppliers are also crucial to effective inventory management, so it pays to cultivate strong supplier relationships. As far as possible, it’s smart to consolidate spend among a smaller number of preferred suppliers to obtain better and more willing cooperation. Look for “steady runners” – attractive business for the supplier – to offset some of the parts that exhibit more volatile demand (and which will require more supplier understanding and assistance).

4.    Customer Behavior is a particular tricky issue to resolve because EMS providers, very understandably, want to provide excellent service to them, and asking for them to provide better forecasts (and live up to them) can be an uncomfortable conversation. But customers can be trained, especially once an EMS has developed a good working relationship with them. Help your customers to understand what problems are caused by random, unexpected demand changes, and try to link those problems to costs that will ultimately be passed on to them.

Adding to these general tools, here are a few specific recommendations:

A.    Focus on cycle time and accuracy in materials handling. You need to be able to count on inventory within hours, not days, of its arrival on your receiving dock; and you need to be able to trust the numbers in your stockroom inventory.

B.    Review the settings in your parts master, to make sure they are consistent and aggressive. It won’t do any good if you have world-class cycle times in materials handling, if your MRP settings still assume a one-week receiving cycle time. The same goes for safety stock levels.

C.    Monitor supplier accuracy and consistency, and take corrective action if a particular supplier can’t achieve high levels of performance.

D.    Ensure that customer demand changes are incorporated into your MRP calculations within hours – not days.

E.    Set up an inbound hub in regions where you have major facilities, or several smaller factories.

F.    Evaluate the volatility of demand (usually characterized via an “XYZ” code) as well as the value of consumption. This is especially important if you carry substantial amounts of finished goods inventory for your customers. [4]

It’s usually possible to make significant initial progress in driving down inventory; but real success comes when an EMS company builds on that foundation and puts well-disciplined processes in place to continually improve its inventory management practices.


[1] See . Some further insights are available at I’m using the larger number for overall “Electronic Products Assembly” which includes Original Design Manufacturing, since ODMs face the same challenges in inventory management as EMS providers, and since many companies pursue both EMS and ODM activity.

[2] Not “activity-based costing” but a method of inventory categorization using a basic application of Pareto’s Principle. See, for instance, or Many companies go beyond the A, B, and C and define a D or [null] category as well, for parts with very low consumption value. Some may also identify a “super-A” category that constitutes the small number of highest-value parts. Often, less than 1% of the total parts count may represent 50% or more of value.

[3] The quality score provides a fair comparison of overall performance, even if individuals manage vastly different complements of parts. It’s calculated as [Sum of target inventory values] / [Sum of actual inventory values].

[4] More info is available on this concept in a separate article, at The idea is to separate the “X” parts (with relatively stable demand), Y parts (somewhat unstable demand), and Z parts (highly volatile, unpredictable demand). Understanding which “A” parts are also “X” (high value, stable demand) vs. “Z” (high value, volatile demand) will help you develop an appropriate strategy for each.