We frequently encounter confusion about the process of setting safety stock levels. This blog hopes to clarify the issue.

Safety stock is a critical component in any system of inventory management. Indeed, some inventory software treats safety stock as the key decision variable in the quest to balance inventory cost against item availability. Unfortunately, that approach is not the best way to strike the balance.

First, realize that safety stock is part of a general equation:

Inventory Target = Average Lead Time Demand + Safety Stock.

Average Lead Time Demand is defined as the average units demanded multiplied by the average replenishment lead time. Example: If daily demand averages 2 units and the average lead time is 7 days, then the average lead time demand is 2 x 7= 14 units. Keeping 14 units on hand suffices to handle typical demand.

But we all know that demand is random, so keeping enough stock on hand to cover the average lead time demand invites stockouts. As we like to say, “The average is not the answer.” The smart answer is to add in some safety stock to accommodate any random spikes in demand. But how much?

There’s the problem. If you try to guesstimate a number for the safety stock, you are on thin ice. How do you know what the “right” number is?  You may think that you don’t have to worry about that because you have a good-enough answer now, but that answer has a sell-by date. Lead times change. So do demand patterns. So do company priorities. That means today’s good answer may become tomorrow’s blunder.

Some companies try to wing it using a crude rule of thumb approach. For instance, they may say something like “Set safety stock at an additional two weeks of average demand.” This approach is seductive: It only needs simple math, and it is clear.  But for the reasons listed in the previous paragraph, it’s foolish. Better to get a good answer than a convenient answer.

You need a principled, objective way to answer the question that takes account of the mathematics of randomness.  More than that, you need an answer that is linked to the key performance indicators (KPI’s) of the system: inventory cost and item availability.

Simple logic gives you some sense of the answer, but it doesn’t provide the number you need. You know that more safety stock increases both cost and availability, while less safety stock decreases both. But without knowing how much those metrics will change if you change the safety stock, you have no way to align the safety stock decision with management’s intent for striking the balance between cost and availability.

Rather than flying blind, you can back into the choice of safety stock by first finding the right choice for inventory target. Once you’ve done that, the safety stock pops out by a simple subtraction:

 Safety Stock = Inventory Target – Average Lead Time Demand.

Often times, companies will state that they don’t carry safety stock because the safety stock field in their ERP system is blank. Nearly always, safety stock is built into the targeted inventory level they have established.  So, using the above formula to “back out” how much safety stock you are building into the plan is quite helpful.  The key is not just to know how much safety stock you are carrying but the link between your inventory target, safety stocks, and its corresponding KPI’s.

For instance, suppose you can tolerate only a 5% chance of stocking out while waiting for replenishment (inventory texts call this interval the “period of risk.”). Software can examine the demand history of each item and work out the odds of stockout based on the thousands of different demand scenarios that can occur during the lead time. Then the right answer for the inventory target is the choice that leads to no more than a 5% stockout risk. Given that target and knowing the average lead time demand, the appropriate safety stock value falls right out by subtraction. You also get to know the average holding, ordering and shortage costs.

That’s what we mean by “backing into the safety stock.” Start with company objectives, determine the appropriate inventory target, then derive the safety stock as the last step. Don’t start with a guess about safety stock and hope for the best.

Share, develop, and manage consensus demand plans to ensure inventory policy matches business strategy.

Smart Inventory Optimization (SIO) creates planning scenarios. SIO starts with a “Live” scenario that shows where you are now. Various team members can create their own scenarios, perhaps dividing the work by product line or sales territory. One decision maker can then merge these scenarios into a consensus plan that becomes the “Goal” scenario, which drives the ERP system’s replenishment planning.

​As the world holds its breathe to see how the new corona virus (2019-nCOV) will play out, we cross our fingers for all those currently in quarantine or under treatment and pray that health authorities around the world will soon get the upper hand.

This short note is about one way your business can develop a plan to adjust to one of the likely fallouts from the virus: sudden increases in the time it takes to get inventory replenishment from suppliers. Supply chains around the world are being disrupted. If this happens to you, how can you react in a systematic way?

Reacting to Longer Lead Times

This is a problem that can be solved using advanced supply chain analytics. Presumably, you may have already used this technology to make good choices for the control parameters used in managing all your inventory items, e.g., values for Min and Max or Reorder Point and Order Quantity. The specific technical question addressed here is how to convert an increase in replenishment lead time to changes in those control parameters.

In general, longer lead times require fatter inventories if you want to maintain a high level of customer service. This general rule translates into larger values of Min and/or Max. How much larger depends critically on what new, longer lead time values will appear and their probabilities of occurring.

While many planning software systems assume a fixed lead time, the reality is that almost all lead times have some degree of randomness. Typically, ignoring that randomness increases stockout risk, so having a good estimate of the probability distribution of lead times is important. In normal times, your transactional data can be used to estimate that relationship. But sudden disruptions like 2019-nCOV create unprecedented situations in which you have to make educated guesses about what new delays you will see and how likely they are. We will assume here that you can imagine some such scenarios and want to figure out how to best respond to them.

An Example using Advanced Software

To illustrate this type of prospective planning, consider a hypothetical example. One item, a spare part, has an established pattern of replenishment lead times, with delays of 5, 10 and 15 days occurring with 15%, 70% and 15% probabilities, respectively. Given this distribution and a random demand averaging one unit every 5 days, values of Min = 5 and Max = 10 do a good job. Figure 1 shows a simulation of 10 years of daily operation under this scenario. Fill rate and service level are high, and stockouts are infrequent.

Now suppose that disruptions in the supply chain create a less favorable distribution of lead time, with a 50:50 mix of 15 and 30 days. Figure 2 shows how badly the current values of Min and Max perform in this new scenario. Fill rate and service level plummet due to frequent stockouts. Operating costs more than triple due to penalties for backorders. Only inventory investment (the average dollar value of stock on the shelf) seems to get better, but this happens only because so often there are backorders with nothing left on the shelf. The shift to longer lead times clearly requires new higher values of Min and Max.

Figure 3 shows how the system performs when the Min is increased from 5 to 10 and the Max from 10 to 15. This change compensates for the longer lead times, restoring the previous high levels of fill rate and service level. Inventory investment is necessarily greater, but operating costs are actually lower than before.

Summary

Changes in normal operating conditions require adjustments in the way inventory items are managed. One such change looming large on this date is the potential impact of the 2019-nCOV Corona virus on supply chains, with anticipated increases in replenishment lead times.

Changes in lead times require changes in inventory control parameters such as Min’s and Max’s. These changes are difficult to make with any confidence using pure guesswork. But with some estimate of the increase in lead times, you can use advanced software to learn how to make these adjustments with some confidence.

This note illustrates this point using simulations of the daily operation of an inventory control system.

Figure 1 Simulation of normal operations using current replenishment lead times, Min and Max

Figure 2 Simulation of abnormal operations using longer lead times and current Min and Max

Figure 3 Simulation of abnormal operations using longer lead times and revised Min and Max

In a previous post, I discussed one of the thornier problems demand planners sometimes face: working with product demand data characterized by what statisticians call skewness — a situation that can necessitate costly inventory investments. This sort of problematic data is found in several different scenarios. In at least one, the combination of intermittent demand and very effective sales promotions, the problem lends itself to an effective solution.

Reviewing terms, recall that “service level” is the probability of not stocking out while waiting for a replenishment order to arrive, while “fill rate” is the percentage of demand that is satisfied immediately from stock. In my previous post, “The Scourge of Skewness”, I pointed out that a certain type of demand distribution, having a “long right tail”, will lead to fill rates that can be much lower than service levels. I also pointed out that sometimes the only way to improve the fill rate is to increase the target service level to an unusually high level, which can be expensive.

In this post, I’ll look at solving the problem in one special case: skewness resulting from effective sales promotions mixed with “intermittent demand”. Intermittent demand has a large proportion of zero values, with nonzero values mixed in at random. Successful sales promotions, obviously positive, have a downside: they can confuse the “demand signal” with spikes in your demand history, and can undermine forecasts and bias safety stock calculations. When intermittent demand and effective sales promotions are the source of your data’s skewness, methods exist to work around the problem to achieve both higher fill rates and more accurate demand forecasts.

How Promotions Increase Skewness

Successful promotions abruptly increase item demand. This creates anomalies, or “outliers”, which contribute to forming a skewed distribution. Knowing when promotions occurred in the past, we can adjust an item’s record of past demand. We produce an alternate demand history as if there had been no promotions, by replacing the outliers with values more representative of the “natural” level of demand. These adjustments reduce demand skewness. Reduced skewness can lead to significant reductions in both expected forecasts and safety stocks, which add together to form reorder points.

Successful promotions are likely to be repeated. When that happens, the promotion effects can be added in to demand forecasts to increase their accuracy. The effect of future promotions on inventory management will be to increase the risk of stockouts, so a sensible response is to work at the operational level to build up temporary supply, in a quantity keyed to the estimated impact of prior promotions on the effected items.

Using Event Modeling to Improve Demand Forecasting

It’s possible to model the impact of like events, and apply this to planned events in the future. Doing so can improve your forecast in two big ways: by projecting the demand jolt you expect from a planned event; and rationalizing the spikes in the past that were caused by events, making your baseline activity more visible and more accurately forecastable. We do a lot of this in SmartForecasts, so allow me to use our experience there to show you what I mean.

Event Modeling entails the following steps:
• Automatically estimating the impact of previous promotions (which is a useful result in itself).
• Adjusting historical demand to statistically remove the effect of promotions.
• Creating promotion-free forecasts.
• Revising the forecasts for any future time periods in which promotions are planned.

We call this this type of analysis “Promo forecasting”. We use the word “promotions” to describe what you do yourself to improve your results. We use “events” to describe what the world does to you, usually to your detriment; examples include strikes, power outages, warehouse fires and other unlucky happenings.

To understand how Event Modeling can help you cope with skewness when doing demand forecasting for high-volume items, consider Figures 1-3.

Figure 1 shows that this item’s demand pattern is clearly seasonal, and the forecast is both seasonal and “tight”, meaning that the forecast uncertainty interval (“margin of error”, shown in cyan lines) is very narrow.

Figure 2 shows an alternative history in which a promotion in June 2014 reversed the usual seasonal low associated with June sales. This demand pattern was forecasted using the Automatic forecasting tournament in SmartForecasts, as in Figure 1. This time, the promotion scrambled the seasonal pattern enough to create an inappropriate non-seasonal forecast, and one that has a much larger margin of error.

Finally, Figure 3 shows how Promo forecasting handles the same promoted scenario, retaining a seasonal forecast and building into the forecast an estimate of the effect of a planned repeat promotion in 2015.

The Case of Intermittent Demand

In Figure 1, the item was a high-volume finished good and the task was demand forecasting. Promo modeling is also useful when dealing with the task of setting safety stocks and reorder points for items with intermittent demand, whether the items are finished goods, components or spare parts. Intermittent demand very often has a skewed distribution that makes it difficult to achieve high item availability with a small investment in inventory.

Figure 4 illustrates the problem that a successful promotion can accidentally create for inventory management. If such a spike arises from the natural, un-promoted demand, then the only way to maintain high fill rates is to provide safety stocks large enough to cope with these random surges. In this case, the big spike in demand of 500 units in February 2013 was the result of a one-time promotion.

Taking Account of Promotions to Improve Inventory Management

Unwittingly treating the spike in the example above as part of the natural demand variability results in a poor fill rate. To achieve a target service level of, say, 95% with a lead time of one month would require a reorder point of 38 units, computed as the sum of an expected forecast over the one month replenishment lead time of 21 units supplemented by a safety stock of 17 units. This investment would result in a disappointing fill rate of only 36%.

However, recognizing that the spike is a one-time promotion and replacing the 500 units with 0 obviously would make a big difference. The reorder point would drop from 38 units to 31 (the sum of an expected demand of 7 units and a safety stock of 24 units) and the fill rate would increase to 94%.

Of course, it is not ok to just throw out inconvenient demand spikes whenever they make life uncomfortable; there has to be a valid “business story” behind the adjustment of historical demand. If the spike is the result of a data processing error, then by all means, fix it. If the spike coincides with a promotion, then replacing the spike with, say, the median demand (often zero, as in this example) will result in a much more sustainable inventory investment that still meets aggressive performance targets. Future promotions of the same type on the same item will require some extra effort to prepare for the temporary surge in demand, but the recommended reorder point will be correct in the long run.

Thomas Willemain, PhD, co-founded Smart Software and currently serves as Senior Vice President for Research. Dr. Willemain also serves as Professor Emeritus of Industrial and Systems Engineering at Rensselear Polytechnic Institute and as a member of the research staff at the Center for Computing Sciences, Institute for Defense Analyses.

Improve Forecast Accuracy, Eliminate Excess Inventory, & Maximize Service Levels

In this Video Dr. Thomas Willemain, co–Founder and SVP Research, talks about forecast-based inventory management policy, also known as MRP logic. This is the fourth in our series on major approaches to managing inventory.  We begin by looking at some very simple and then more robust models of inventory dynamics that help us determine how much to order or manufacture and when. We then consider how to calculate lead time and account for lead time variability. Tom concludes by describing the importance of safety stock, it’s role in properly buffering against demand and supply uncertainty, and how best to calculate it.