Your Privacy Matters

Parts of this website use cookies and content from third-party providers. More details can be read in our privacy policy, which also describes how you can amend your preferences at any time.

DeclineAccept

Skip to main content

When British mathematician and computer scientist Clive Humby coined the expression “data is the new oil” in 2006, he was making a point not just about the value of data, but also about the need to refine, distil and make by-products from it to make it useful to drive business innovation.¹ While Professor Humby’s primary focus has been consumer data, his observations are equally true of enterprise data.

Major enterprises generate no small amount of data. And within that data resides key information to allow you to optimize your inventories. But what does that end-to-end data analytics process look like? And how can organisations improve it to further optimize their inventories?

A major weakness in the relevant data sets for inventory analytics is their format and their accuracy. The data you need is usually deep within Enterprise Resource Planning (ERP) systems or similar. This means that it is already challenging to extract the data in a format conducive to analytics. A surprising number of organisations consider even getting good visibility of what inventory they have a victory. Furthermore, data accuracy problems within those data sets are legion. Here are some of the most common causes of data inaccuracy concerned:

• Transactional data is subject to booking errors, especially in non-discrete manufacturing but also elsewhere
• The relevant master data structures can be complex and misunderstood. To take just one example, there are often multiple lead time fields in ERP, with limited documentation as to what each should contain, leading to incorrect and inconsistent use
• Many master data fields require manual input, leading to errors, gaps and data that is no longer up to date
• Some master data field values are subject to considerable variation

When you consider that for a large enterprise you may be dealing with millions of transactions and hundreds or thousands of items, getting your data into sufficient shape is a daunting task in itself. Where data is contained within an enclosed system, such as a warehouse management system, it is possible, using technology like RFiD, to build and maintain very accurate data sets that then enable all kinds of optimizations. Unfortunately, when it comes to the wider supply chain this is not the case. Too many of the critical data points for inventory optimization rely on externalities for you to simply use raw ERP data without significant curation.

Let us imagine that you have a perfectly refined data set. Now you need to distil it in a way that will allow you to analyse your inventories. Here we can differentiate between descriptive and prescriptive inventory analytics (what inventory you have and what inventory you should have). Descriptive inventory analytics requires minimal statistical analysis. It provides basic inventory visibility and classification, breaking inventory down by type and over time. This might include things like aging stock reports, ABCXYZ analysis, and so on. It is a very useful overview, but doesn’t tell you how close to optimal your inventories are.

In order to progress into prescriptive inventory analytics, you need to calculate what an optimal level would be. Here, you have a choice between an approach that primarily uses inventory science (equations) and an approach that primarily uses modelling by way of simulations. In principle, it should be possible to achieve the same results with either approach, however modelling is significantly more time- and power-hungry than an equation-based approach. Modelling is particularly useful when you don’t know what precisely what you’re looking for. Equations are much more efficient when you do know precisely what you’re looking for.

With either approach, the chief challenge is the multiple variables and sensitivities involved in an optimization calculation. Especially if you’re dealing with a complex operation with many different types of products and processes, one-size fits all is highly unlikely to give you equally accurate or useful target inventory values for everything. While it may seem that modelling is free from some of the complexities of equations, in reality each model is highly dependent on the starting parameters you feed it and experimenting with different values is again extremely time consuming. Usually only the model builder themselves understands the model well.

Before undertaking the first two steps, you should ideally have a clear idea of what you are working towards. Ultimately, the questions most people are trying to answer about their inventories, are:

1. How much do I really need?
2. What should I do differently?

Statistical analysis is a great way of answering the first question and a vital step in answering the second. We are going to look at the second question in Step 4 (driving business innovation), but in the design of our inventory analytics by-product we should ensure that it answers the first question well and that it does so in a way that makes it useful for the next step.

Working out how much inventory you need comes down to either modelling or calculating the minimum inventory you need to hit target service levels. But when you have done this, how do you know if you’ve got it right? The first option is to put blind trust in your analysis. A preferable option is to test the conclusions of your analysis. However, this runs into a number of practical challenges, the first of which is change. Inventories and the variables that influence them are constantly changing. By the time you have tested your analytics and discovered problems, it is difficult to ascertain whether your issue is change or your analysis itself.

In reality, single value answers are of limited value. Instead, good inventory analytics should measure a whole range of properties for every single item to allow multiple segmentations. There is also a balance to find between precision and utility. There is almost no end to the number of adjustments you could make to an optimization calculation based on different properties of each item. But your analysis then becomes much more difficult to understand and is liable to fail at the next step. On the other hand, oversimplify your model and your analysis will only be valid for a subsection of your inventory.

For the by-product of your inventory analytics to be valuable it must find the right balance between precision and utility. It must also be:

• Built using sufficiently refined data (see step 1)
• Distilled using an appropriately robust process (see step 2)
• Rigorously tested to ensure validity
• Be sufficiently scalable and automated that it can be refreshed frequently and easily for all of your inventories

It takes significant time and multiple iterations to create a by-product that is genuinely fit for purpose. Avoid manual processes or over-reliance on individuals otherwise your accuracy and sustainability are heavily compromised.

Inventory analytics is ultimately only useful if it allows you to better optimize your inventories. Too many organizations put a lot of effort into analysing their inventories but then stumble at the final stage of getting operations to take advantage of the analysis. Why is this final step so difficult?

The simple answer is complexity and change. There are many different factors that affect your inventory levels and the gap between actual and optimal inventories. Let us explore some of the major ones:

1. A mismatch between how optimization calculations are derived and how planning is done on the ground. A good optimization calculation will factor in variability in quite a different way to how human brains normally do it and quite likely in a different way to any planning tools or methods that you might be using. An optimization calculation will take all available data and work out the very best levels. Most planning processes, on the other hand, look primarily at the forecast and then factor in some kind of safety stock based on expected variability. However, the planning process does not necessarily use the same statistics or assumptions as the optimization calculations. This means it can be very difficult to understand what needs to change to achieve optimal levels
2. In the planning process, change is constant. Planners spend a lot of time dealing with changes in demand, in production schedules and in expected deliveries. Today’s inventories are normally the result of decisions taken weeks if not months ago. Untangling everything that has changed in the intervening time is not only laborious and difficult, but sometimes simply impossible if forecasts and schedules are constantly overwritten. This can lead organisations (often wrongly) to assume that forecast accuracy is their only or biggest issue: “we must have expected we’d need this”
3. Over-reliance on systems, tools and models within the planning process. Companies spend a lot of money on technology that promises to optimize inventory. But, even if the technology itself is fit for purpose, if the data within it, its parameter settings, or simply the way it is being used are not correct then it will not properly help to optimize your inventories. An assumption that the technology is right, however, leads to complacency. Where it disagrees with your inventory analytics, it is extremely difficult to understand what is driving the difference
4. Insufficiently refined raw data. However smart your analytics is, if your raw data is not sufficiently refined then it will produce results that your planning team know to be wrong, destroying confidence in the whole of the analysis
5. Organisational and behavioural constraints. Sometimes inventories are sub-optimal because of conflicting incentives elsewhere in the organisation (for example, production trying to maximise output) or because planners are unwilling to ignore forecasts they know from experience to be wrong. This type of effect is highly prevalent but invisible in inventory analytics other than the results.

So what is the solution? How can organisations better use inventory analytics to deliver sustainable improvements to their inventory levels? For now, let’s assume that you have got a perfect by-product, with perfectly refined data and an optimally distilled analysis.

The first use of good inventory analytics is in reactive measures. Whether or not you understand the root causes of your inventory imbalances, you can at least see and address them. This might mean postponing or cancelling orders or production. This is often the method companies choose when they have a temporary or cyclical drive to reduce inventories. Depending on your starting position and lead times, it can be possible to reduce inventory by 10% in just a few months while improving or maintaining service levels.

Such a rapid improvement in your cash position is significant and very valuable. However, it doesn’t deliver sustainable inventory performance if done as a one-off exercise. Many inventory reduction programmes are successful in the short term only for inventory levels to rise again after the focus comes off. One way to avoid this is simply to keep iterating the process. Analyse your inventories twice a year or once a quarter and continue making reactive adjustments. If your inventory analytics can be done quickly, accurately and at scale across your business at a reasonable cost, this is not a bad solution. At nVentic, we have developed such a solution, which we call the Inventory Evaluation.

Ideally, rather than postponing or cancelling unnecessary orders, you would avoid raising them in the first place. To do this, you need to align the cadence of your analytics with your planning cycles such that your planners have useful decision support. But in this case, you need to turn your analytics around very quickly and present it to planners in a very simple format. Again, nVentic has developed a solution for this too which we call the Inventory Projector.

However, good inventory analytics allows you to go further still by providing insights into the root causes of inventory imbalances. If your planning processes and parameters were optimised in the first place then you wouldn’t need so much reactive work. Here we need to differentiate between insights that you can glean directly from your analytics and insights that you can only develop by considering your processes in parallel. To take an example as seemingly simple as safety stock, the optimal amount will vary depending not just on the distribution of your demand (which your analytics should tell you) and on your lead time (the accuracy of which you can only ascertain with further effort) but also on the planning approach you follow. If you plan deterministically but calculate safety stock using a replenishment model then you won’t have optimal inventory levels.

In short, the best inventory analytics imaginable will not allow you to sustainably optimise your inventories unless you also put sufficient effort into matching your systems, processes and KPI’s to what the analytics is telling you. Too many organisations have teams of talented analysts producing good analytics that their operations then fail to take advantage of. One solution to this is to use an automated inventory analytics solution like nVentic’s Inventory Evaluation. This allows you to switch your analysts’ focus from producing analytics to interpreting and applying the analytics. Only in this way can you bridge the gap between the analytics and operations.

The other obvious advantages of using a pre-packaged and automated solution like nVentic’s are that it has been rigorously tested with a variety of industries over more than 10 years, separately both measures and mitigates your data quality issues, works directly with raw data from your ERP, uses the best scientific methods and comes with support to help you interpret your data.

A huge amount of improvement is possible with advanced inventory analytics. Our clients regularly see improvements of 20% or more. But the complex nature of supply chain planning for most big organisations means that they need to put much more effort into the interpretation and application of the analytics. Many planning solutions promise inventory optimization but fail to deliver it, not just for some of the reasons outlined above, but also due to the fact that optimization calculations are by their nature too precise when you consider the uncertainties in the underlying data. Humans are still much better at dealing with situations that are unclear.

Progress in big data analytics carries on apace and you should definitely put effort into inventory analytics. But why not change your focus from production to application?

Contact nVentic to discuss how to make a game changing switch to your inventory optimization efforts.

Notes

1. https://www.sheffield.ac.uk/cs/people/academic-visitors/clive-humby

To see the latest version of this DIO benchmark, see our latest white paper: nVentic Big Pharma inventory trends benchmarking report 2025

We ended last year’s benchmark report with a question. Would 2023 be the year that the seemingly inexorable upward rise in big pharma inventory levels abates? The answer, emphatically, is no.

Without one-off effects,¹ in DIO terms median inventory grew a further 6.1% year on year, with mean inventory growing 5.1%. Total big pharma inventory grew from $160bn to $175bn. And this in a year where revenues and cost of sales were down 4%.

Estimated overall inventory write offs held steady at about 4% of cost of sales.² However, there were significant write offs of Covid-19-related vaccines and treatments, most notably Pfizer’s eye-watering $6.2bn write off of Paxlovid and Comirnaty.

Days Inventory Outstanding (DIO) = inventory value/(cost of sales/365). See also the technical notes at the end of this article.

Overall, despite a good crop of new medicines coming to market, financially 2023 wasn’t a great year for the industry. There were notable blockbuster successes, such as MSD’s Keytruda. Growth in demand for weight-loss treatments carried on apace, although with greater competition as more medicines enter the market. However, the overall lacklustre sales saw many companies forced to cut costs, with layoffs across the sector. The full impact of the Inflation Reduction Act on pricing still remains to be seen but is also driving cost focus.

We update the companies included each year, so you can’t compare the numbers from one year’s report to another. However, in the seven years we have been producing this report, DIO has increased every year but two. DIO remained roughly flat one year and showed a small decrease the other. The other five years have shown increases. Looking at it another way, in 2017, median DIO was around 180 days, equivalent to approximately 6 months of demand. In 2023 median for those companies who were in the benchmark in 2017 was over 220 days, comfortably more than 7 months in demand.

In short, the industry’s inventory binge does not appear to be over. As we write every year, there are good reasons why pharmaceutical companies carry a lot of inventory. These include security of supply for products that are critical to patients, long lead times and regulatory restrictions. However, there are also bad ones, such as high gross margins and a lack of focus on inventory optimization.

One reason for the increase in big pharma inventory in 2023 was the playing out of a classic bullwhip. Concerns of shortages throughout the value chain in previous years had led many to increase inventories, leading to overstocking. Producers of API’s noted a slow-down in sales, as many pharmaceutical manufacturers had swollen their raw material supplies in recent years to ensure supply. For the companies in our benchmark, on average raw materials inventories are almost double what they were before the pandemic. Meanwhile the pharmaceutical manufacturers themselves reported overstocking at their customers as a reason for suppressed demand.

Given the vital importance of their products, no one would want the pharmaceutical manufacturers to run short of essential medicines. This consideration naturally features very highly in their approach to inventory. Are the current levels of excess and obsolescence in big pharma inventory not simply a necessary evil? And is it not a price worth paying?

While a certain amount of waste is necessary to maintain very high service levels for products with very short shelf lives, most medicines have a shelf life sufficiently long for this not to be a significant factor on its own. When a product has a shelf life of several months, this is certainly the case.

Another way of looking at it is from a purely financial perspective. There are a variety of companies in the benchmark: generics manufacturers, research-driven pharmaceutical manufacturers, diversified groups, contract manufacturers, but the average gross margin in 2023 was 65%. The gross margin is so large because R&D accounts for such a large proportion of the true cost of doing business for a lot of the firms, but EBITDA is also very healthy in the sector. In this context, scrapping 4% of your inventory each year equates to little more than 1% of revenue. This is a very small consideration compared to the rewards for, say, successfully bringing a new blockbuster to market.

Inventory is not a cost you want to minimize. It is a strategic asset you want to optimize. Medicine shortages continue to be a global problem.³ As we highlighted in last year’s report, perhaps counterintuitively, shortages and overstocks are not inversely linked. Yes, if you only have one product then having too much of it means you can’t simultaneously not have enough. But given the huge number of different medicines and dosage forms, it is obviously possible to have too much of some and not enough of others and this is the case in practice. Manufacturing capacity, raw materials and cash are all finite resources. Overproduction in one area can and does cause shortages in others.

Avoiding inventory optimization to avoid risking shortages is the wrong way of thinking about it. You should reduce unneeded inventories so that shortages can be better avoided. This is not to say that pharma supply chain teams are not already exerting themselves to better achieve this balance. Of course they are, this is the very essence of supply chain. But as we look at another year’s inexorable rise in big pharma inventory levels, we have to conclude that there is still much room for improvement.

If you have any questions about this report, or would like help improving your organisation’s inventory levels (whether you feature in the benchmark or not), please contact us.

Technical Notes

1. “One-off effects” include the change of companies in the benchmark as well as major acquisitions or divestitures. When we compare DIO numbers from one year to the next, we exclude companies with this type of one-off effect so that they do not distort the underlying trend. Of course, a lot of these companies are making acquisitions or divestitures on a fairly regular basis, so we only exclude those where a significant impact on inventory levels is observable in either of the two years compared. The major one-off effects in 2023 are as follows:

• Bausch Health recorded a fair value step up on $162m of XIIDRA inventory acquired from Novartis. This should be amortized over 2 years, of which $20 was amortized through Cost of Sales in 2023.
• Biogen recorded a fair value step up on $1.3bn of inventory acquired through the purchase of Reata. This should be amortized over 3 years, with 31.5m amortized in 2023. Biogen have finished writing off Aduhelm.
• Pfizer recorded a fair value step up on $4.1bn of inventory acquired through the purchase of Seagen , $672m of inventory acquired through the purchase of GBT and $817m of inventory acquired through the purchase of Biohaven. Pfizer wrote off $6.2bn of unmarketable Paxlovid and Comirnaty (which also affected the DIO numbers by swelling Cost of Sales). In addition, Pfizer lost $209m of inventory from the tornado which hit its Rocky Mount facility.
• Amgen recorded a fair value step up on $5.025bn of inventory acquired through the purchase of Horizon Therapeutics.
• AstraZeneca recorded a $114m impact on Cost of Sales from the unwinding of fair value from the acquisition of Alexion.

At the time of publishing, Boehringer Ingelheim had not released their Finance Report for 2023. We have replicated their 2022 figures as a placeholder and will update the benchmark once it becomes available. Boehringer Ingelheim do not report Cost of Sales so we have used an estimate as with previous years.

2. For estimates of big pharma inventory write offs, see our article “Inventory write offs in pharmaceutical manufacturing”.

3. See, for instance, the European Association of Hospital Pharmacists report, this summary from C&EN on shortages in the US, or this research briefing from the UK parliament. Of course, there are multiple reasons why medicines run short, but a lack of free capacity to react quickly to shortages is one.

Inventory optimization is a much-used phrase, but what exactly is it, how do you achieve it and do you in fact even need it? As our mission as a company is to help organisations get value from inventory optimization, this is a topic very close to our heart.

In this article we consider inventory optimization to be the application of mathematical techniques to define ideal inventory levels. Such theoretical models are not easy to apply in practice. The idea that a piece of software can allow you to optimize your inventories simply and automatically is attractive but at present far from possible.

We believe that the biggest value of inventory optimization techniques is in showing you how much better your inventories could be, while the correct interpretation and application of that information is likely to require changes to how you manage your inventory.

The benefits of overcoming these challenges are substantial. Inventory optimization allows you to free up working capital, reduce shortages and obsolescence, reduce your environmental footprint and increase flow through your supply chain, usually by significant amounts.

You will often see any improvement to inventory management described as inventory optimization, but this is not strictly speaking accurate.

In mathematics, optimization means selecting the best option from a set of alternatives. When talking about inventory, this comes down to determining the best amount of inventory to hold to meet your objectives. Inventory optimization mathematically seeks to find the sweet spot between having too much and not enough.

This is a simple concept to grasp in principle, but not to apply in practise. In order to calculate an optimum level mathematically, you need to factor in a number of different considerations, some of which are themselves difficult to quantify.

Let’s say that you want to optimize on purely economic principles – what will make you the most profit? To do this you need to be able to quantify, on the one hand, the true holding cost of inventory and, on the other hand, the cost of shortages, which might include backorder costs, expediting costs, the cost of lost sales and lost market share, even the cost of damage to your brand. Noting that these costs are entwined with each other. It’s not that this is impossible to do, but it is laborious and somewhat imprecise.

In fact, it would be fair to say that few people, if any, actually try to calculate this. Instead, normal practice is to set a target service level, either implicitly or explicitly, and then try to minimize inventories while staying at or above that service level. This is a good, pragmatic move, that ties down one of the more difficult parameters to set. However, even with a pre-defined target service level, it is still challenging to calculate optimum inventory levels.

When it comes to improving inventories, we differentiate between optimizing, which involves working with the givens, and improvements based on changing the givens. What are the givens? They are the parameters you need as inputs to your inventory optimization calculations. We have already considered target service level as a given, but this is just one of many such parameters.

When working with the givens, you accept them as fixed parameters and optimize your inventory levels on that basis. Which is not to say that the givens can’t be improved, but you need such parameters defined to even attempt an optimization calculation.

How can you optimize your inventories?

Inventory optimization comes down to defining two things for each item: the optimal order quantity/batch size (Q) and the optimal safety stock (SS), noting that the latter might be zero or even negative. But there is no one way to do this for all items you stock, since the method chosen will itself vary depending on a number of factors. Let us start by taking a couple of simple examples.

1. You have a product that is made to order. Whenever a customer orders one, you have 10 weeks to make it. It takes one week to make and 4 weeks to source the parts. Everything you need is at or above your minimum order quantities (MOQ’s). You have no inbound supply variability or capacity constraints.
2. You have a product that is made to stock. Demand is relatively stable, with a limited amount of variability, as is supply. You check inventory levels of finished goods continuously. You schedule production to avoid stock outs. You try to maintain raw materials inventory to be able to produce when needed.

If anything in bold changes, you will need to factor it into your optimization calculation, along with the other factors shown in Figure 1 above.

Different types of inventory require not just different optimization calculations, but also different planning methodologies.

For the first example above, we can consider a deterministic planning model. This is appropriate when you know what demand is going to be. In this instance, you can set your order and batch sizes to exactly what is required. In principle, you shouldn’t need safety stock, although you might need to have some safety stock to protect against variation in supply, such as late deliveries or quality problems in production.

For the second example, we can consider a replenishment model. This is appropriate for make-to-stock finished goods and buy-to-stock raw materials where demand is not too intermittent but displays variability. In this case, the EOQ model or one of its variants should be used to calculate optimal order/batch size. Safety stock can be calculated using the target service level, the standard deviations of demand and lead time, and the lead time itself.

However, this replenishment model, which is based on probably the most widely known text-book set of equations, has important sensitivities. The most commonly found version of the safety stock equation, for instance, assumes that demand is normally distributed, that average demand does not vary, and it also breaks down when the variability is very high. You will sometimes find people saying that the equations “don’t work”, but this is often just their understanding of the equations running into some of these barriers which they either don’t understand or at any rate don’t know how to work around.

Note that different models might work best at different points in your supply chain. For instance, you might choose a replenishment model for finished goods and a deterministic model for raw materials. To really optimize your inventories, you are likely to need to employ different strategies for different types of inventory.

And these two examples are far from representing the only types of stock you might carry. For instance, you may have make-to-order items where the customer promise time is shorter than the total supply lead time. You may have spare parts with highly sporadic demand but which need to be available on demand. And so on. Each different type of inventory needs an appropriate optimization approach.

In order to optimize your inventories, you need to take all the factors above into consideration for every item. And so far, we’re just talking about mathematical optimization. In addition – and no less a challenge – you have to consider what processes, systems and people are in place and how they would need to change in order to deliver and maintain the optimization.

In our experience, the biggest gap is often in the people themselves and how they usually work. The knowledge of optimization techniques and how to take advantage of them is a rare commodity amongst material planners, whereas data scientists do not necessarily have the requisite understanding of the practicalities of the business to enable practical application of their models. Technology is sometimes seen as a way around this but we are far from this being a satisfactory solution yet.

Do you need inventory optimization?

From the brief summary above, it should be evident that inventory optimization is not simple. In fact, it is closer to the truth to call it impossible than to call it simple, but this is not an all or nothing situation. Given the complexities outlined, as well as the fact that your optimal inventory levels are constantly changing, fully optimized inventory is not an achievable goal. But the application of optimization techniques is incredibly valuable. You don’t have to win an Olympic gold medal to get value from learning how to swim. Moreover, many of the biggest benefits from inventory optimization come from the first steps.

In the absence of inventory optimization, inventory management is flying blind. You are searching for the sweet spot between too little and too much inventory without knowing where it is. Think of it as trying to find a black mouse in a completely dark room. With optimization techniques, you are switching the light on. It is still difficult to catch the mouse, but at least you can see it.

It is of course possible to manage your inventories without trying to optimize them. Especially in times of abundant cheap cash this is what many organisations choose to do. But without the analytics of inventory optimization you have no real way of knowing how well you’re doing. Or rather, your view becomes lopsided. Your service level performance tells you if you have enough, but what is telling you if you have too much? Top-down aggregate measures like cover, inventory turns or DIO give you ratios that indicate whether you’re getting leaner over time. You can also use them for quick benchmarking exercises, although actually they tell you nothing about how much is optimal.

Where organisations do not use optimization techniques, there is a strong tendency to apply “one size fits all” approaches to different types of inventory: for instance, setting target safety stock levels for all raw materials at 2 weeks’ average demand. This could only be optimal if all raw materials have identical variability of demand and supply plus identical lead times. This approach means there is a strong correlation between shortages and inventory levels long before you reach an optimized level. Because inventory optimization by necessity works at an individual item level, it allows you in aggregate to reduce inventories and shortages simultaneously.

From our own empirical experience of working with multiple organisations, it is normally the case that inventory optimization alone (so taking the givens as given) allows net inventory reductions of 20-50% while maintaining or improving service levels. This improvement takes time and effort to achieve, but double-digit improvements are normally achievable in the first year. We can think of no other option that delivers such large improvements so quickly in such a risk-free fashion.

How do you take advantage of inventory optimization?

The good news is that you do not need to solve all of the technical problems at once. Some of the biggest benefits come from the simplest steps.

To take advantage of inventory optimization you need two things. Firstly, the ability to run the optimization calculations. Secondly, and no less importantly, the ability to operationalize the findings of your analytics. You will only sustainably improve inventory levels if your organisation can use the insights generated day to day.

The biggest value in inventory optimization analytics comes from the visibility it gives you into the biggest opportunities for improvement. It highlights things that are not necessarily obvious or intuitive. In the first instance, a calculation using a number of approximations is normally good enough. Once you have a good grasp of this, you can go further and seek to remove or refine the input approximations one at a time, noting that you will never remove all of them entirely.

Because the analytical challenge is high, you need specialist skills to do it properly. But those skills are not necessarily the same ones required to make changes to how your business actually works. Too much emphasis on analytics, or a siloed mentality, can leave a large gap between theory and practice. The analytics itself is only one part of the picture. In addition, you need to build the people, processes and systems to take advantage of it.

Inventory optimization is not something you can implement and then forget about. Rather, it is an ongoing process which needs constant attention. All of the input parameters to your optimization calculations are in a state of almost constant change. The right process controls need to be in place to optimize your inventories, especially the closer you get to optimal.

If done well, inventory optimization can transform how well your supply chain functions. However, this requires a clear understanding of what it can do and how to use it. Year one of a well-designed programme can easily deliver double digit improvements from some of the easiest steps. Those who want to get much closer to optimization can deliver that type of improvement for many years in succession.

Why write a guide to inventory for CFO’s? Because the CFO is usually a necessary sponsor of any working capital programme.

What impact would a 25% reduction in inventory have on your business? Well:

• It would make a nice difference to your balance sheet. It would reduce working capital and overall capital employed. It would reduce the amount written off for obsolescence each year
• It would make a nice difference to your income statement. Typically inventory holding costs (including finance costs) can add up to 25% of the value of inventory held each year
• You would have some cash available for more interesting uses than sitting in a warehouse: capital investment, acquisitions or equity buy-back for instance
• And it would make a nice difference to the ratios your investors care about. Both Return on capital employed and return on equity will improve, as will Earnings per share (EPS), especially if you choose to use the liberated cash for a share buyback

Suggesting an improvement in working capital to most CFO’s is a case of preaching to the choir. To illustrate the impact of a 25% reduction in inventory on key financials we have included an example in the appendix below. For a €2.5bn turnover company with €400m in inventory such as the one we take in our example, it improves ROCE by 2%, EPS by 3% and ROE by over 4%. These percentages will heavily depend on your circumstances; the point is that the difference is enough to care about. This guide to inventory for CFO’s will look deeper at the potential and typical challenges.

The case for holding inventory

Most CFO’s would sign up for the scale of improvement above if they believed it was achievable. But is less inventory automatically a good thing? Well not if it is at the expense of customer service or sales. The whole purpose of inventory after all is to provide a buffer between demand and your ability to supply. If profit margins are high, and raw materials markets unpredictable, the temptation to hold plenty of inventory can be strong. Our contention is that most firms have significantly more than they need, whilst still managing to run short of some items.

A hundred years ago, inventory was considered a good thing. The attitude was “pile them high, sell them cheap”. Assets were a source of pride. But increasingly, leanness is considered a virtue. In times of constant innovation and change, the risk of obsolescence is high and almost all products can seem ephemeral. Increases in on-demand production and the ease of finding alternatives via online services is only intensifying this trend. Online retailers can actually achieve negative working capital. Whilst this isn’t the case in most other industries, we do see that the ability to be agile and to hold inventories much closer to ideal levels is increasingly going to be a source of competitive advantage.

To get an approximate sense of how close to the mark you are, look at your Days Inventory Outstanding (DIO). This measure takes your total inventory value and divides it by cost of sales days (total cost of sales/365). The number you get will not tell you very much per se – different businesses will have significantly different inventory needs. However, comparison with your peers should be informative, and the figure will allow you to sense check it yourself.

If your DIO is equivalent to 80 days, for instance, as in the simplified example in the appendix below, that means on average you are holding over 2.6 months of stock. Does it take you that long to source, produce and ship your products, even allowing some spare for demand fluctuation? It might do, if you have long lead times, a lengthy production process and high demand variability, but these are the types of question to ask your Operations teams. Bear in mind that with an average of 2.6 months of stock, you will have significantly more of some items. Reducing DIO to 60 days, again following our example, still gives you 2 months’ inventory on average.

This is one of the headaches for a finance executive. Supply chains are different. Depending on your production strategy (make to order vs make to stock), the seasonality and variability of your demand, the lead times in your supply chains and the economics of your target service levels, the “right” amount of inventory for you can vary enormously. Appropriate policies will also vary depending on your industry and the nature of your supply chain.

Plus there are of course other levers to improve working capital, which may well seem much simpler to pull. Reducing accounts receivable is mostly about internal process discipline. Even increasing accounts payable, whilst often labour intensive, is usually just a case of lengthening payment terms with your suppliers. And increasingly, supply chain financing options exist to help improve this balance depending on your own particular working capital needs. By comparison, inventory can seem a very complex undertaking. It is one which we know many firms have tried and failed to address in the past.

Between having an intuition that you might be holding excess inventory and actually being able to see exactly what you should have, let alone do something about it, there is quite a gap. Many firms have tried and failed to seriously get a handle on inventory. Very few who have tight control of inventories and a strong understanding of the different levers which contribute to it. But why is this so?

In our experience, at the heart of the problem lies data. A reasonably large firm might easily have thousands of different items between finished goods, work in progress and raw materials. Each transaction should be captured in your systems, and stocks reconciled with physical stocks at least once per year. With this data, you have everything you need to analyse your inventory and identify optimal levels BUT the data is hard to extract, manipulate and then calculate.

Inventory planning software is highly sensitive to relatively minor input adjustments which are frequently poorly understood. As a result, even mathematically sound planning tools fail to do the job of optimizing inventory and don’t really give you any insight into how you are performing overall and what can be done to improve the situation.

From a lack of clear and accessible data stems a number of other challenges: inability to correctly identify the root causes of sub-optimal stock levels, and a politicisation of inventory. Most major inventory decisions are based on opinion and perception rather than objective analysis. All of these factors will tend towards excessive stock being accumulated.

Cutting through the data complexity of inventory for CFO’s

How nice it would be to have the following picture available for every item you stock, as well as a roll-up of overall optimization potential:

A = actual historical stock on hand. B = average of A. C = target stock on hand. D = average of C. E = excessive stock, or difference between actuals and target.

This graph shows a familiar basic supply chain concept. The blue “sawtooth” line (C) represents what you should have had on stock. The green line floating above it (A) represents what you actually had. Fairly simple to plot a course of action from this picture.

Of course, the simplified model upon which this diagram is based needs to be enhanced with probabilistic considerations to account for all kinds of variability. You also need to understand demand patterns, which is a science in itself. This raises the bar somewhat on the mathematical front. It also demands segmentation along ABCXYZ lines to identify appropriate policies for different types of stock. But with this type of granularity you have a perfect springboard to drive root cause analysis item by item.

We meet many executives who understand the principles at play here, but precious few who have access to the ability to extract large volumes of data (many millions of transactions) and perform the appropriate mathematical analysis on them. We have personal experience of consultants being brought in to do the heavy lifting. A team can be buried away for 2 to 3 months crunching the numbers, which drives cost and is highly prone to human error. Or high-level assumptions are made to avoid going into this level of detail, but then the business case is less robust and the roadmap of what to attack is less clear and usually focuses on macro-processes such as planning and forecasting.

Consultant-led projects normally do succeed in taking out inventory in the short term. If nothing else the executive focus their presence guarantees will deliver this. However, it is less clear that this benefit is sustainable, and this engenders reluctance to try this route again.

We also find that inventory optimization software has a tendency not to deliver on its promise. Whilst the better of these tools are built on similar mathematical principles to the more basic of our own tools, they are mostly focused on deriving target inventory levels and they are highly sensitive to inaccuracies in input data such as lead times or demand forecasts. We find many instances where companies are simply overwriting the targets generated by the software since this is an easier fix than understanding and addressing the underlying issues. It is easy to spend a lot of money on inventory optimization software without getting much value from it.

What we have done at nVentic is to automate the data analytics required to truly understand what’s going wrong in your inventory. We provide technology that gives you this insight at speed and at scale. This provides unparalleled insight into inventory for CFO’s.

In addition, we firmly believe that the only way firms will develop sustainably optimized inventories is by building capability in their internal teams. Our preferred way of working is therefore to partner with a client team to deliver change. nVentic provides the analytics and the expertise, and we help guide your team through a process of understanding the root causes and addressing them over a period of months and in some cases years. We know that many companies out there, particularly in the manufacturing sector, have the potential to reduce inventories by much more than 25% whilst actually improving service levels.

Getting there is likely to take more than one year though. Your teams need to learn to walk before they can run, and whilst you most likely already have a number of strong inventory people, they need to be able to bring the rest of the organization with them. 25% is typically a reasonable target for year 1 however.

Is it time for you to get your inventories in shape?

To learn more about optimizing inventories, contact us for an initial discussion.

Appendix

Illustration of the impact of a reduction in inventory for CFO’s*

Let’s say your business turns over €2.5bn each year, with cost of sales of €1.85bn and net income of €250m. On the balance sheet you have €400m inventory, €600m in receivables and €700m in cash. In addition, you have non-current assets of €1.2bn, giving total assets of €2.9bn. On the liabilities side you have €900m in current liabilities and €800m in non-current liabilities. With €1.7bn liabilities you therefore have €1.2bn in equity.

Your capital employed is €2bn. Using this highly simplified model, you therefore have ROCE of 12.5% and ROE of 20.8%.

Let’s say you reduce your inventory from €400m to €300m, a 25% reduction. Since studies have shown the holding costs of inventory to be approximately 25% of the value of inventory each year, you will at least in theory increase net income by €25m. (In fact, the P&L impact will almost certainly be less than this, since holding costs include both variable and fixed costs, and absorption can be a challenge, especially in the first year of a programme, but for simplicity here we will use the full 25%. See also our article on holding costs.)

Your ROCE will thus improve to 13.8% and your ROE to 22.9%. If you just transform the reduced inventory into cash you will not have affected your net working capital overall, although your acid-test ratio will have improved. You will also most likely need to write down less inventory at the end of the year. Maybe an improvement from €20m to €15m written down. But with the example we have taken, liquidity ratios are not a concern and cash reserves are generous. More likely that with a return on equity of 20%+ there is a strong case for further investment or a share buyback to boost earnings per share (EPS). If price per share is €20 and there were 150m shares on average to begin with, 5m shares could be bought back and earnings per share increase from 1.833 to 1.896, an improvement of nearly 3.5%.

We can summarize the numbers as follows (assuming the share buyback):

* The “Before” figures for this illustration are all taken (approximately) from a real company, the one major change we made being to reduce their actual starting inventory from ~€700m to €400m.

Inventory science is a field well represented in the academic world for over a hundred years. From at least Harris (1913) onwards,¹ work has been done to explore how to calculate optimum inventory levels. But how close are theory and practice in actual inventory management.

If you look at what is going on in organizations today from a practical perspective, you will find surprisingly little of that research being applied. In this article we’re going to explore why this is, what impact the failure to apply inventory optimization principles has, and finally what we’re doing at nVentic to help bridge the gap.²

Mind the gap (between theory and practice)

So why is inventory science so popular with academics but under-developed in practice? In a word, complexity. Academics love a good challenge. And there is sufficient complexity in inventory to make it an attractive discipline.

Outside of academia, however, complexity is usually not so welcome. If something is difficult or onerous to do, there is a strong temptation to make do without it. Let us have a look at why inventory science is so complex.

The image below is familiar to most students of supply chain:

On the basis of a few key parameters, we can determine how much should be ordered, and when, to maintain optimum inventory levels. We can also determine, based on a few other parameters, how much safety stock to keep.

Why is there a gap between theory and practice?

This is a simple model, and not too difficult to follow. However, in our experience, there are few organizations using it to any great extent, if at all. There are a number of barriers to using it in practice:

1. Some of the parameters are hard to quantify. To build the model above for any SKU we need to be able to quantify re-order costs, holding costs and shortage costs (the forth cost parameter, unit cost, is usually easier to quantify)
2. We need to layer variability into the model. Demand is not uniform or constant, and it may not even demonstrate normal distribution. The same may well be true of lead time. This demands some more advanced mathematics
3. Everything in this model needs to be done SKU by SKU. Organizations often have thousands of SKU’s, so the data burden alone is often a roadblock
4. The mathematical model is too precise. The underlying equations will deliver results in fractions, but in reality, organizations need to deal with minimum order quantities, stock controls that aren’t instant, and so on

We will come onto what nVentic does to counter these problems below. First, let’s consider what impact a failure to apply inventory optimization principles will have.

Pile them high, sell them cheap

When Harris was writing, in the early 20th century, inventory was considered a good thing. But over the last 50 years in particular, leanness has come to be considered a virtue. In a world where customers expect greater choice, and constant product innovation, and are only a click away from finding it online, holding too much inventory can be very risky. Holding inventory costs money and ties up working capital.

Of course, you can improve your inventory situation without applying optimization techniques. Measures like shortening lead times and using consignment stock make it easier to operate with less inventory on your books. But you still have to decide how much stock to hold, how much to order and how often. If organizations are not using optimization models, how do they do this? How are theory and practice different?

If an optimization model is not used at all, the alternative is often a simplified concept of how many days’ or weeks’ stock you need to hold. Some ready reckoning will factor in average demand, demand peaks, perishability if appropriate, lead times and convenient (for logistics, production or purchasing) lot quantities. Although even doing this requires quite a lot of manual effort and data manipulation which isn’t always easy.

Underused optimization technology

Accordingly, some organizations help themselves by using planning software which will leverage at least some of the mathematics we describe above. However, as we have already noted, these models are very sensitive to the input data, so using tools can be prone to error. We have seen many instances of software being overwritten or ignored because early experience with it led to undesirable outcomes.

Our view is that most organizations are not doing much in terms of inventory optimization. The 2017 REL Europe Working Capital Survey suggests €350bn is tied up in excess inventory in the biggest 1000 companies in Europe alone. And what is harder to quantify because not publicly reported, but no doubt an even bigger concern to many organizations, are the sales lost due to shortages.

That £350bn of tied up working capital is a wasted opportunity, since it could be put to more interesting uses, such as acquisitions, paying down debt, or equity buy-backs. Moreover, as the holding cost of inventory is typically in the region of 20%, that’s also €70bn of unnecessary cost. And that’s before factoring in the impact of lost sales. If companies haven’t addressed this sooner it is surely not only because of the difficulty, but also because most of their competitors haven’t either. Yet it represents a major, largely untapped, opportunity to gain competitive advantage. Between theory and practice a lot of cash is left on the table.

Build a bridge and get over it

At nVentic, our mission is to close the gap between the academic theory and practical application of inventory science. So what do we do to enable the bridging of theory and practice in inventory optimization?

Firstly, we have built analytical tools which will do all of the calculations for you automatically based on historical data, removing the opportunity for manual error. Secondly, we work with clients to actually optimize their inventories. We believe that the tools in isolation have limited value but by working collaboratively with our clients we help them not just to realize the benefits identified, but also to build internal capability for sustainable improvement.

We come to inventory optimization with a strong command of the mathematics, which we have embedded in our own analytical tools, but moreover with many years of helping organizations make concrete improvements in inventory management.

The nVentic approach

In terms of dealing with the 4 barriers to applying the mathematical model in practice, we apply the following approach:

1. (Parameters which are hard to quantify.) Here the key is approximation. Sensitivity analysis shows that modest changes in these parameters will not make significant differences to the outcome. We can advise clients on what to use here based on experience
2. (Factoring in variability.) This is vital and we use advanced algorithms to deal with it. However, there are also practical ways to simplify it if necessary in the short term by looking at maximum values
3. (Dealing with large data sets.) We have developed tools which take raw data from your system(s), automatically transform them and perform over 100 mathematical routines on them. The output is an overall quantification of potential, and the segmentation of your SKU’s by value/volume and variability (ABCXYZ analysis). In addition, our tools allow you to drill down by SKU: historical stock evolution, key parameters and deltas (how much stock did you have compared to how much you optimally should have had?) We can perform this analysis in a matter of days even on very large data sets
4. (The mathematical model is too precise.) We deal with this in two ways. Firstly, we automatically round values up depending on known constraints such as packing sizes or minimum order quantities. Secondly, we recommend an incremental approach to making improvements. Rather than jumping straight to the calculated target, aim to build in a little spare buffer in the first instance. Then work towards optimal levels incrementally, focusing on the most valuable SKU’s

We find that even taking a conservative approach, significant double-digit improvements are possible for most organizations in the first year, with plenty more to come in subsequent years. The transparency that our tools provide allows excess stock holdings to be quickly drained. And then we work with clients to interpret the data and implement the underlying changes which are necessary to embed optimized inventory management. In this way they build internal capability to make year on year improvements with an increasing command of the subject.

Conclusions

Bridging the gap between theory and practice in inventory management is something we believe in passionately. To truly optimize inventory takes time and effort, but significant improvements are possible quickly and, thanks to the technology we have developed, easily. If you would like to build a bridge in your organization, reach out to us for a conversation today.

Notes

1. Ford W. Harris, “How Many Parts to Make at Once.” Factory, The Magazine of Management 10 (2), February 1913, 135-136, 152. And “How Much Stock to Keep on Hand.” Factory, The Magazine of Management 10 (3), March 1913, 240-241, 281-284.
2. Another topic that we’re not considering here is why Finance departments tend to abandon the topic of inventory management to Operations given how strategic working capital is. This in itself leads to inventory being neglected as a strategic asset, with more emphasis on the tactical topic of inventory control. See also Linda G. Sprague and Marc J. Sardy, “Inventory Management: Some Surprising news about classical views on inventory and some non-classical responses to traditional practice”, Inventory Management – non-classical views, ed. M.Y. Jaber, 2009, Chapter 2.

The opportunity to improve working capital is large across the economy. This is an ongoing unrealised potential. Macro studies from consultancies like The Hackett Group and PwC calculate the opportunity in the trillions of Euros.¹ There are many interesting points to note in these reports, of which we will highlight three:

1. There is a wide spread in performance between companies. Hackett finds that upper quartile companies convert cash 7 times faster than the median
2. Sustainability is a challenge, with very few companies demonstrating consistent improvements year on year. Both studies highlight this
3. The only working capital lever consistently applied over recent years is Days Payable Outstanding (DPO). PwC finds that companies have increased pressure on their suppliers rather than managing their own assets better. But new payment terms legislation in several countries is limiting how much more can be achieved through DPO

All of this resonates with what we at nVentic observe as inventory optimization specialists. The opportunity to improve working capital through inventory optimization is great, but few organizations have achieved sustainably high performance in this area.

Why should this remain an unrealised potential? Hackett highlights what it believes to be a reason why inventory has unrealised potential:

“Inventories can often be a… complex area to drive working capital optimization due to competing cross-functional objectives (cash/cost/service) and the ability to easily identify core improvement drivers from within. […] Generally, no one function can drive change without the participation and collaboration of the others, and the process re-engineering and optimization requirements needed to release cash internally can sometimes be politically overwhelming if the exact causes of excess inventories have not been determined.” (Hackett, 2018 US Working Capital Survey)

Inventory can indeed be a complex area to improve, and not just due to the cross-functional challenges involved. Inventory optimization even as a paper exercise requires a strong understanding of the key levers, some of which are complex in themselves and most of which are in a state of flux: the data challenge alone can be daunting. And even once you’ve determined the best course of action you still have to take the rest of the organization with you.

Our mission at nVentic is to help companies cut a path through the complexity. We have developed advanced diagnostic tools which automatically identify optimum inventory levels based on real data. But equally importantly we know how to handle multiple variables simultaneously and produce a clear and pragmatic way forward. For the truth of the matter is that the vast majority of companies can deliver significant double-digit improvements in their inventory position even before they address cross-functional issues.

Let us dig into why you (probably) haven’t optimized your inventory:

1. Do you know what good looks like?

How many organizations really know much inventory they should have? While it is practically impossible, given all of the variables in play, to say how much inventory an organization should have precisely, and recognizing that this theoretical number changes all the time, there is still enormous benefit in calculating what you need as accurately as possible. In the absence of a statistically derived target, organizations may rely purely on top-down or operationally-derived targets. Each has its weaknesses.

Top-down, an organization may simply aim to increase its inventory turns each year. All very well, but how do you set a target? For instance, your inventory might currently be turning over 12 times each year. What is your target for next year going to be? 13 times? 14 times? The point is, without calculating an optimized inventory policy bottom-up, item by item, you can’t know what is possible or desirable. It could be that 12 is already optimal and that by targeting 13 or 14 you are simply going to increase shortages. Or, much more likely, it could be that optimal is more like 20, in which case a target of 13 or 14 is undershooting.

You could derive a target by benchmarking your peers, and this is generally not a bad starting place, but even here you must take care. Optimal inventory levels are very heavily influenced by variables such as lead times, capacity utilization and target service levels, which may well differ greatly even between peers. And how do you know if even your best performing peer is that close to optimal?

At an operational, planning level, organizations are most likely already deriving their own targets as they try to ensure they have enough to meet expected demand. Because this is done on an ongoing basis the implied targets may never be formalised as such. This type of target is the province of Operations and rarely gets visibility at a board level. Or put another way, the target is used for planning and not for performance management. It is used to drive immediate action (order more, produce less, etc.). It is not a goal to work systematically towards.

Related to this question of what good looks like is what you measure:

2. What KPI’s do you use to assess inventory performance?

KPI’s drive behaviours, so the KPI’s you use to measure inventory performance will have an important impact on your overall performance. We can divide inventory KPI’s into two broad groups:

1. Measures of whether you have enough inventory
2. Measures of whether you have too much inventory

You can judge whether you have enough inventory by measuring how successfully you meet customer demand. The vast majority of organizations use one or more measures of service level, whether it be fill rate, cycle service level, shortages, back orders, lost sales, ready rate, out of stock time, or similar.

Having enough inventory is usually a higher priority for most organizations than avoiding having too much. We find that most organizations lack KPI’s to avoid overstocking, unless it is a high level one like inventory turns. And this is already a driver of underperformance. If you don’t have some kind of balance in your KPI set – so that your teams are trying to keep inventories within a range, rather than just one side of a minimum – you will almost inevitably carry too much inventory across the board. The occasional, or not so occasional, shortage only heightens the perception that shortages are the real risk.

Knowing what good looks like is a prerequisite to having a good set of balanced inventory KPI’s.

Also important is how those KPI’s are used – who is measured by them, who sees them and how targets are set and balanced. If different parts of your organization (say Sales and Operations) are using KPI’s in direct conflict with each other without a robust Sales, Inventory & Operations Planning (SI&OP) process in place, it can quickly turn into a turf war rather than a collaborative effort to optimize overall business outcomes. Remember that KPI’s drive behaviours!

3. Who has inventory targets as part of their personal goals?

This flows from the KPI discussion. Mostly likely your inventory planners do. But how about your general managers? A planner can only do so much without the alignment of the rest of the organization. Of course, you don’t want everyone bogged down with the minutiae of inventory metrics, but how about charging 20% of the inventory value (a reasonable approximation of holding costs) back to the P&L of each general manager? See what difference that makes to inventory levels!

4. Do you consider inventory optimization to be risky?

If you hear people talk about reducing your inventory by 25% do you immediately think of shortages?

We have encountered few executives who wouldn’t welcome reduced working capital in principle. But if organizations consistently fail to take full advantage of this opportunity, one of the main reasons is surely the perceived risk. The whole purpose of inventory after all is to enable you to supply your customers with as much of your product as economically desirable. Especially when the cost of capital is low and profitability is strong, why risk running short when you can just carry a little more inventory? Yet inventory optimization and inventory reduction are not synonymous. Good inventory optimization initiatives simultaneously reduce inventory levels and shortages.

This counter-intuitive phenomenon is due to the individuality of every single item you hold in inventory. Each one has its own unique profile in terms of demand, demand variability, lead time variability, perishability, normality of distribution, and so on and so forth. What we invariably find when we analyse organizational data sets is that while there is too much of some (and usually most) items, there is not enough of a smaller number. Inventory optimization addresses both of these issues. This is why you can simultaneously reduce your overall inventory levels while increasing inventories of some items and therefore reducing shortages. This is one of the few occasions in life where you really can have your cake and eat it!

5. Are you spending money on inventory optimization tools but are still unsure what good looks like or what value you have really got from them?

We are not opposed to the many inventory optimization tools on the market. On the contrary, we believe there are several good ones out there. But here is the rub. Are they being used, and are they being used to their full potential? Plus do the people using them understand their limitations and sensitivities? We come across many organizations with licenses to advanced inventory optimization tools that they barely use, if at all. In many cases trials were made with the tools but problems were encountered so they are switched off and old and trusted manual methods are reinstated. The possible causes of confusion are numerous, but to list a few common examples:

• Do your tools require a single lead time, and do you know what to do if your lead time varies?
• Are your tools dependent on you feeding them a forecast and are you using unconstrainted sales forecasts which invariably fail to materialize?
• Are your tools reliant on data in your ERP system? Do you know which data and how accurate and up-to-date they are?
• Do your tools offer you alternative settings that you find hard to choose between because you don’t really understand what each means or what difference it will make?
• Do your tools produce values that you know from all practical experience to be incorrect, but you can’t tell why?
• Do your tools assume normal distribution of demand? If so, do you know how to test for normality and what to do with other types of statistical distribution?

In short, inventory optimization tools and statistical models more generally can be immensely valuable, but only if used correctly. Used incorrectly they can be positively detrimental. Because the expertise to get the most out of these tools is rare, they tend not to deliver on their promise. This is (often) not so much a fault in the tools themselves as in how they are used.

6. Do you manage to reduce inventories from time to time when you have a big push on working capital, but find the levels always seem to rise again?

Most things in life they improve given attention, but when neglected they soon revert to a lower level. Yet there seems to be something about inventory which makes it particularly prone to this phenomenon. Why does sustainability seem to be such a challenge? We believe there are two main reasons.

Firstly, inventory reduction can be quick and simple to deliver. You just need to buy and/or produce less! Thus, when companies have a sudden need to free up working capital a strong top-down mandate is usually sufficient. But done crudely, this creates problems, since shortages usually follow. After the initial squeeze, therefore, inventories return to their previous levels as operations revert to their normal way of functioning. Nothing really changed other than a temporary prioritisation.

Secondly, thorough approaches require a certain level of expertise and a concerted effort. When there is a big push on inventories, consultants may be brought in and additional internal efforts also allocated. But this increase in resource is temporary and this in itself creates problems with sustainability. Any deep statistical analysis conducted has the disadvantage of being static.

Rather like inventory optimization tools, one-off projects are not a silver bullet. To deliver sustainable improvements in inventory you have to invest in internal capabilities, frequently refresh analysis and maintain top-down pressure.

7. Do you find inventory a difficult topic? Would you rather focus on less complex things?

It makes sense to do easy things that will make a positive difference before turning to more difficult ones. The business case for optimizing inventories is normally extremely strong – a reduction in working capital and holding costs, combined with an improvement in service level, all delivered at a very strong return on investment. If most organizations have not done more inventory optimization, it is because it is difficult.

While not underestimating the human or political difficulties involved, we believe that what differentiates inventory optimization from many other business change initiatives is the difficulty working out how much inventory you really need, which needs to be calculated on a regular basis. But without this essential first step, done really thoroughly, all of your other efforts will be sub-optimal. You won’t be able to set the right targets, using the right KPI’s, allocated to the right people. You will be at risk of running short and potentially spending money on initiatives without being able to tell how successful they really are.

Yet we believe this difficulty is also one of the reasons why inventory optimization is a great topic to address. Precisely because inventory optimization has unrealised potential, significant double-digit improvements are possible in most cases.

To return to one of the Hackett metrics with which we opened: top quartile companies convert cash 7 times faster than median companies. On how many metrics do you have this kind of advantage over your competitors? Surely this is worth overcoming some difficulties to achieve?!

Conclusions

There are many reasons why organizations fail to maintain optimal inventories. We have explored 7 common ones here, of which knowing what good looks like is the most important. Based on this insight, nVentic has developed diagnostic tools which quickly and accurately show clients how much inventory they should hold. We then work with them to embed sustainable inventory capabilities in their teams.

Until the macro reports start showing a marked and sustained decrease in inventories, our work will not be complete.

If you would like to discuss how to define what good looks like for your organization and take your inventory optimization capabilities to the next level, contact us.

Notes

1. For the most recent editions of these reports at the time of publication, see The Hackett Group, 2018 Europe Working Capital Survey and 2018 US Working Capital Survey, and PwC, Navigating Uncertainty, PwC’s annual global Working Capital Study 2018/19. Each of these documents looks at a different selection of companies and each finds a potential of at least 1 trillion Euros or Dollars for the selection they look at.

How much does inventory really cost to hold? Inventory holding cost may not seem like a strategic lever. In fact, it is one of the key ways in which Finance departments can help improve working capital.

In 2019, a commonly encountered belief is that holding cost is low because interest rates are low. This is partly true, but misleading. In this white paper we will examine why, and why it matters.

One reason why organisations may hold too much inventory is that they underestimate the true cost of holding inventory. In this white paper we will also examine why this is an issue, what a realistic holding cost is and how it should be used.

What is inventory holding cost?

Holding cost is the cost of holding inventory and is expressed as a percentage of unit costs. If an item costs €100,000 to produce or buy, and it costs €20,000 a year to hold, its holding cost is said to be 20%.

Holding cost is made up of the cost of capital tied up in inventory, plus the operational cost of holding the inventory.

Importantly, holding cost is related to but not the same as the P&L (EBIT) impact of increasing or decreasing inventory. Holding cost is the total cost of holding inventory and is used in inventory management for calculating batch sizes and deciding on advance or discount purchases. It is greater than the P&L impact of changes in inventory for a number of reasons. Firstly, EBIT does not include the cost of capital, but secondly, holding cost includes a number of costs which you cannot switch on and off at will: warehouses, employment, loans, etc. Holding cost includes P&L items, but also balance sheet effects.

It is a mistake to ignore these “fixed” costs. If you believe inventory to be expensive to hold, you will try to hold very little. If you believe inventory to be cheap to hold, you will tend to hold a great deal. But smart organisations should avoid this. Why fill up a building with inventory when you might have better uses for the space, such as adding a production line, or subletting part of the building? And once the building is more or less full, a perception that all that inventory is needed can lead to sub-optimal longer-term decisions, such as adding a warehouse.

Remember there is no such thing as fixed costs, just variable costs with longer or shorter time horizons.

How much is inventory holding cost?

Each organisation will have its own unique holding cost depending on its cost of capital and its operational cost structure. Holding cost can be difficult to estimate accurately, but it is worth putting some effort into generating an approximate figure.¹

The operational cost typically comprises the following categories:

1. Storage costs. Both external and internal warehouses, including, for instance
   1. (notional) rental costs
   2. Maintenance
   3. Temperature control
   4. Security
   5. Insurance
2. Obsolescence. This includes sub-categories such as:
   1. Scrap (the total value of the inventory written off)
   2. Rework costs
   3. Discount costs (selling items at below cost)
3. Internal transfer costs. Moving inventory around between sites
4. Handling costs including, in particular, stock counting/stock control and especially those causing production stoppages
5. Management costs. The time spent by planning teams controlling existing inventory, dealing with obsolescence etc.
6. Overheads. All of the above requires IT, HR, Finance and other corporate time and expense

It is important to keep in mind the difference between total inventory holding cost and P&L impact. You may (rightly) feel that you need a storage warehouse, but this does not mean it has no cost. You will also need to make some estimates. If, for instance, you own a facility where half the space is filled with production lines and half the space is storage for inventory, then half the relevant costs for the facility should be included in inventory holding cost. Or if you have a team who are responsible for ordering, planning and controlling inventory, you should only include their time spent controlling existing inventory in holding cost. And so on.

Then consider the cost of capital. This should be the Weighted Average Cost of Capital (WACC). It is true that, writing in 2019, the global cost of debt is at historically low levels² and this in turn has led to relatively low WACC for asset-intensive companies, but this can be misleading. For a highly indebted company, the fact that WACC is low does not mean it can easily issue more debt. For such a company it would be undesirable to make decisions on the assumption of a low holding cost since it will lead to more capital being tied up in inventory. The opposite should be the case – these companies should do all they can to free cash from inventory and reduce their dependence on debt.

In fact, there is a compelling argument to ignore the calculated holding cost altogether and consider the opportunity cost instead. Look at the Internal Rate of Return (IRR) threshold for internal business cases approval. What is the highest IRR project that was rejected in the last 12-24 months? This can be considered the opportunity cost of the cash you have tied up in inventory and is a useful sense check that you are using a reasonable holding cost rate. The opportunity cost can easily be lower than your inventory holding cost, but if it is higher there is a strong strategic case to use it as your holding cost.

Why does it matter – what is holding cost used for?

Holding cost is an important factor in the inventory levels organisations target and ultimately hold. This is because of the decisions which are made on the basis of assumed holding cost. To take the two most important instances:

• You can calculate optimal procurement order sizes by offsetting re-order costs and volume discounts against holding cost. This directly affects raw materials inventory levels
• You can calculate optimal production batch sizes by offsetting set up costs against holding cost. This directly affects finished goods inventory levels

Note in these two examples we say “can”. In many cases organizations make these decisions with no explicit holding cost. Instead, they might produce “lot for lot”. Or they might simply choose the batch/order sizes that minimize production/acquisition costs within the constraints of what is possible in terms of capacity and reasonable in terms of expected demand. If you find a year’s worth of a particular raw material in your warehouse because Procurement found a great deal on it, this effect is in evidence.

These two uses of holding cost also feed into other strategic decisions, such as capital investment decisions, make vs buy decisions and even strategic supply chain design.

It is easy to underestimate the importance of holding cost for working capital. In inventory optimization initiatives the focus is often on safety stock levels, but order/batch size is another key lever. Let me illustrate this with a simplified example:

Demand for a product is 100 units per month. Production is in 400 unit batches every 4 months. This means on average (net of safety stock), 200 units are on stock over the course of a year.

Now halve the batch size to 200 units produced every 2 months. Average stock holding (again, net of safety stock) over the year is now only 100 units.

So what should you do?

Holding cost remains difficult to quantify precisely. Some approximations are necessary and it is usually not worth putting too much effort into it. But the final point to consider is the behavioural aspect. Operations departments have a constant focus on EBIT and only an occasional focus on cash. EBIT focus is a good thing, but can lead to an unconscious prioritisation of EBIT over cash. A key way to counter this is to use a realistic measure of holding cost in calculations of batch sizes and TCO calculations for procurement.³

We recommend you start by looking at how you are using holding cost today and ensure that you are using a holding cost that is not too low. 20% is normally a good starting point, or WACC plus 10% as an absolute minimum. We then recommend trying to calculate your holding cost bottom up until you are confident you are in the range. Avoid having individual plants or business units do this themselves, since it leads to divergent results and much duplication of effort. Instead, Finance should establish a standard holding cost centrally and then mandate its use globally. Multi-national corporations should find that holding cost varies from country to country in line with interest rates. Sometimes high inflation countries should use a higher holding cost than the global standard.

From a behavioural perspective, if Operations has no working capital targets on their dashboard, there will be a strong tendency to hold too much inventory for the reasons we have set out in this paper. Fixing this is also important. But setting holding cost centrally is a strong lever that Finance can exert to free up working capital. Are you getting the most you could out of it?

If you would like to discuss how to take your inventory optimization capabilities to the next level, contact us.

Notes

1. An internet search for inventory holding cost can quickly generate an approximate percentage for rule of thumb calculations. For instance, Investopedia suggests 20-30% although it does not cite a source for this statistic.

A textbook from 1993 reviews estimates of holding costs in academic studies between 1951 and 1990 (Lambert, D.M. and Stock, J.R., Strategic Logistics Management (third edition), Irwin (1993), page 366). All but one is within the range of 20-29% and while that one exception goes as low as 12%, that is only as part of a range from 12-34%.

Interest rates are indeed very low in 2019, even within the perspective of the last 100 years. When Harris (1913) first derived the EOQ model he took holding costs to be 20% at a time when interest rates (in the USA) were at 4.5%, so only 2% more than 2019.

The CSCMP (Council of Supply Chain Management Professionals) commissions a yearly state of logistics report which includes a benchmark of inventory holding rates over the past 10 years in the US. The most recent figure, for 2018, was 18%.

2. The European Central Bank at the time of writing in 2019 was actually offering negative interest rates on deposits (and 0% on loans). Of course, most companies cannot borrow at ECB base rates, but costs of borrowing are historically low, as is WACC, which factors in the cost of equity. Some very big companies, based on what they report in their annual reports, have WACC in the low single digits.

A benchmark of US firms from the start of 2019 found the average WACC for 6000 US non-financial firms was 8.22%. A KPMG study from 2018 focusing just on 276 firms in the DACH region, including 26 of the DAX30, is more or less identical (7% average WACC when including financial firms). Note that operational costs can easily add another 10-20% on top of WACC.

3. This often adds additional impetus to reducing EBIT effects too. An organisation that complacently uses large batch sizes with an assumption of low holding cost (or no sight of holding cost), when confronted with a requirement to factor in a holding cost of 20% or more, will often urgently have to address topics like SMED to avoid a negative impact on productivity.

If you asked most supply chain professionals what super power they would most like, the ability to see the future would probably come fairly high up the list. Uncertainty – of demand, but also of lead times, failure rates, pricing, exchange rates, and many other variables – creates a high percentage of the headaches which supply chain professionals have to deal with on a daily basis. This makes forecasting a critical focus.

But how much would 20-20 foresight really help? Ask Oedipus’s father! He got 100% forecast accuracy from the Oracle at Delphi, but little good it did him.¹ Amongst the many forecasting methodologies used by modern professionals, consulting high priestesses in Greece doesn’t normally feature too highly. But the Oedipus story illustrates an important limitation of forecasts: when you don’t have all the details, knowing what the future holds is not as useful as you think. This principle is very important in inventory management.

Limitation of forecasting for inventory management

For the non-specialist, it can be tempting to think that forecasting is the challenge with inventory optimization. If only we knew what demand is going to be, goes the thinking, we could always have just enough inventory. There are two major dangers with this argument:

1. Forecasts are almost always wrong due to the inherent unpredictability of the future
2. Too much emphasis on forecasts can blind you to the importance of other factors. Even with a perfect forecast, you can easily get your inventory levels wrong

Let me illustrate this second point.

Product A Daily demand: 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2

Product B Daily demand: 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 60

In both instances the monthly demand is 60 units, with an average daily demand of 2 units. If your forecast was for monthly demand of 60 units, well done, your forecast accuracy is 100%!

Now think about how much stock you need to fulfil all orders in each case. Daily demand for product A is uniform – you need no safety stock. Each day you ship 2 units and produce 2 units for the next day. On average you hold 2 units. Product B is very different. In order to meet orders, assuming you don’t know which day the order(s) might fall, you would need to carry 60 units every day. That’s 30 times as much stock as for product A. Quite a difference, considering forecast accuracy was perfect and total demand was identical in both instances.²

It is said that forecasts are always wrong. The further into the future they look and the more detailed they are, the more wrong they are. Knowing this, forecasters normally refresh their forecasts regularly and avoid too much granularity – grouping demand into weekly or monthly “buckets”. But as our example has shown, even a perfect forecast of aggregate demand has limited usefulness without an equally good understanding of what variability to expect.

Rather than focussing too much on forecasting, inventory managers need to give adequate consideration to other factors, such as demand variability, lead time, target service levels, and so on. However, this is not to say that you don’t need to give proper attention to demand. It remains one of the most important inputs to inventory calculations and is probably the most likely to change. But to consider demand you don’t have to turn to forecasts.

Forecasting is a substantial discipline in its own right and one which we’re not going to explore in detail here. At nVentic, our approach when calculating optimal inventory levels is to start by using actual daily demand instead of any forecast. There are a number of benefits to doing this:

1. You already know your actuals, so you don’t need to put effort into generating a forecast
2. Actuals are in their natural form and not aggregated into buckets that hide the actuals
3. You can use actuals to measure variability as well as average demand
4. Actuals are incontrovertible – you don’t need to spend time convincing other people how realistic they are
5. Actuals are immutable – over time more data become available, but the actuals for any given time period will not change. Forecasts, however, are constantly overwritten and it is easy to lose track of what they were at the time plans were put together, especially since lead times vary by product

There are certainly cases where actuals are inadequate: for instance, with new product launches, where there are no actuals; or sporadic spare parts, where mean time between failures (MTBF) models may well be more useful. And there are events that planners need to be aware of and allow for that may not be visible in the actuals (sales promotions, patent cliffs, seasonality, etc.) as well as longer term product cycles which may be evident in the actuals (sales growth, decline, etc.).

We are not suggesting that using actuals is always better than using forecasts, or that you shouldn’t put efforts into improving forecast accuracy. Our approach is to start by using actual demand to calculate optimal inventory levels and then factor in the forecast if (and only if) you are confident it will be more accurate or more beneficial than the actuals.³

Conclusions

In inventory management, it is essential to understand the limitations of any forecast and not neglect the other key levers. Properly factoring demand variability into the approach is often a largely untapped area of potential. Shortening lead times takes some effort to achieve, but allows organisations to be more responsive and to carry less stock. And developing explicit service level targets and building them into your inventory approach is a very valuable way of balancing the opposing requirements of high service levels and low cost in an objective way. So don’t neglect forecasting, but don’t let it blind you to other key inventory optimization levers either.

If you follow this approach, we predict you will be pleasantly surprised by the results.

Notes

1. King Laius consulted the oracle at Delphi as to whether his wife, Queen Jocasta, would ever have children. The oracle told him that any boy born from their union would end up killing his father. Therefore, when a boy, Oedipus, was born, they left him out on a hillside to die. Oedipus was spared but due to the sequence of events grew up ignorant of who his true parents were. In a classic example of a self-fulfilling prophecy, Oedipus kills Laius precisely because he doesn’t know who he is. (Which one can assume would not have happened had Laius not heard the prophecy in the first place!)
2. This example is deliberately simplistic. In reality a majority of items are likely to display more normally distributed demand, but these two extreme examples illustrate the point with a minimum of statistics.
3. How do you determine if the forecast you have been using is better or worse than actuals? In statistical terms, if the Mean Absolute Deviation (MAD) of your forecast vs your actuals is greater than the MAD of your actuals, you would be mad to use the forecast!