Customer as Teacher

Our customers are great teachers who have always helped us bridge the gap between textbook theory and practical application of forecasting and demand planning. Our latest bit of schooling concerns “blanket orders” and how to account for them as part of the demand planning process. 

Expanding the Inventory Theory Textbook

Textbook inventory theory focuses on the three most used replenishment policies: (1) Periodic review order-up-to policy, designated (T, S) in the books (2) Continuous review policy with fixed order quantity, designated (R, Q) and (3) Continuous review order-up-to policy, designated (s, S) but usually called “Min/Max.” Our customers have pointed out that their actual ordering process often includes frequent use of “blanket orders.” This blog focuses on how to incorporate blanket orders into the demand planning process and details how to adjust stocking targets accordingly.

Demand Planning with Blanket Orders is Different

Blanket orders are contracts with suppliers for fixed replenishment quantities arriving at fixed intervals. For example, you might agree with your supplier to receive 20 units every 7 days via a blanket order rather than 60 to 90 units every 28 days under the Periodic Review policy. Blanket orders contrast even more with the Continuous Review policies, under which both order schedules and order quantities are random.  In general, it is efficient to build flexibility into the restocking process so that you order only what you need and only order when you need it. By that standard, Min/Max should make the most sense and blanket policies should make the least sense.

The Case for Blanket Policies

However, while efficiency is important, it is never the only consideration. One of our customers, let’s call them Company X, explained the appeal of blanket policies in their circumstances. Company X makes high-performance parts for motorcycles and ATV’s. They turn raw steel into cool things.  But they must deal with the steel. Steel is expensive. Steel is bulky and heavy. Steel is not something conjured overnight on a special-order basis. The inventory manager at Company X does not want to place large but random-sized orders at random times. He does not want to baby-sit a mountain of steel. His suppliers do not want to receive orders for random quantities at random times. And Company X prefers to spread out its payments. The result: Blanket orders.

The Fatal Flaw in Blanket Policies

For Company X, blanket orders are intended to even out replenishment buys and avoid unwieldy buildups of piles of steel before they are ready for use. But the logic behind continuous review inventory policies still applies. Surges in demand, otherwise welcome, will occur and can create stockouts. Likewise, pauses in demand can create excess demand. As time goes on, it becomes clear that a blanket policy has a fatal flaw: only if the blanket orders exactly match the average demand can they avoid runaway inventory in either direction, up or down. In practice, it will be impossible to exactly match average demand. Furthermore, average demand is a moving target and can drift up or down.

How to Incorporate Blanket Orders when Demand Planning 

A blanket policy does have advantages, but rigidity is its Achilles heel.  Demand planners will often improvise by adjusting future orders to handle changes in demand but this doesn’t scale across thousands of items.  To make the inventory replenishment policy robust against randomness in demand, we suggest a hybrid policy that begins with blanket orders but retains flexibility to automatically (not manually) order additional supply on an as-need basis. Supplementing the blanket policy with a Min/Max backup provides for adjustments without manual intervention. This combination will capture some of the advantages of blanket orders while protecting customer service and avoiding runaway inventory.

Designing a demand planning process that accounts for blanket orders properly requires choice of four control parameters. Two parameters are the fixed size and fixed timing of the blanket policy. Two more are the values of Min and Max. This leaves the inventory manager facing a four-dimensional optimization problem.  Advanced inventory optimization software will make it possible to evaluate choices for the values of the four parameters and to support negotiations with suppliers when crafting blanket orders.

This blog provides an overview of this topic written for non-experts. It

• explains why you might want to read this blog.
• lists the various types of “machine maintenance.”
• explains what “probabilistic modeling” is.
• describes models for predicting downtime.
• explains what these models can do for you.

Importance of Downtime

If you manufacture things for sale, you need machines to make those things. If your machines are up and running, you have a fighting chance to make money. If your machines are down, you lose opportunities to make money. Since downtime is so fundamental, it is worth some investment of money and thought to minimize downtime. By thought I mean probability math, since machine downtime is inherently a random phenomenon. Probability models can guide maintenance policies.

Machine Maintenance Policies

Maintenance is your defense against downtime. There are multiple types of maintenance policies, ranging from “Do nothing and wait for failure” to sophisticated analytic approaches involving sensors and probability models of failure.

A useful list of maintenance policies is:

• Sitting back and wait for trouble, then sitting around some more wondering what to do when trouble inevitably happens. This is as foolish as it sounds.
• Same as above except you prepare for the failure to minimize downtime, e.g., stockpiling spare parts.
• Periodically checking for impending trouble coupled with interventions such as lubricating moving parts or replacing worn parts.
• Basing the timing of maintenance on data about machine condition rather than relying on a fixed schedule; requires ongoing data collection and analysis. This is called condition-based maintenance.
• Using data on machine condition more aggressively by converting it into predictions of failure time and suggestions for steps to take to delay failure. This is called predictive maintenance.

The last three types of maintenance rely on probability math to establish a maintenance schedule, or determine when data on machine condition call for intervention, or calculate when failure might occur and how best to postpone it.

Probability Models of Machine Failure

How long a machine will run before it fails is a random variable. So is the time it will spend down. Probability theory is the part of math that deals with random variables. Random variables are described by their probability distributions, e.g., what is the chance that the machine will run for 100 hours before it goes down? 200 hours? Or, equivalently, what is the chance that the machine is still working after 100 hours or 200 hours?

A sub-field called “reliability theory” answers this type of question and addresses related concepts like Mean Time Before Failure (MTBF), which is a shorthand summary of the information encoded in the probability distribution of time before failure.

Figures 1 shows data on the time before failure of air conditioning units. This type of plot depicts the cumulative probability distribution and shows the chance that a unit will have failed after some amount of time has elapsed. Figure 2 shows a reliability function, plotting the same type of information in an inverse format, i.e., depicting the chance that a unit is still functioning after some amount of time has elapsed.

In Figure 1, the blue tick marks next to the x-axis show the times at which individual air conditioners were observed to fail; this is the basic data. The black curve shows the cumulative proportion of units failed over time. The red curve is a mathematical approximation to the black curve – in this case an exponential distribution. The plots show that about 80 percent of the units will fail before 100 hours of operation.

Figure 1 Cumulative distribution function of uptime for air conditioners

Probability models can be applied to an individual part or component or subsystem, to a collection of related parts (e.g., “the hydraulic system”), or to an entire machine. Any of these can be described by the probability distribution of the time before they fail.

Figure 2 shows the reliability function of six subsystems in a machine for digging tunnels. The plot shows that the most reliable subsystem is the cutting arms and the least reliable is the water subsystem. The reliability of the entire system could be approximated by multiplying all six curves (because for the system as a whole to work, every subsystem must be functioning), which would result in a very short interval before something goes wrong.

Figure 2 Examples of probability distributions of subsystems in a tunneling machine

Various factors influence the distribution of the time before failure. Investing in better parts will prolong system life. So will investing in redundancy. So will replacing used pars with new.

Once a probability distribution is available, it can be used to answer any number of what-if questions, as illustrated below in the section on Benefits of Models.

Approaches to Modeling Machine Reliability

Probability models can describe either the most basic units, such as individual system components (Figure 2), or collections of basic units, such as entire machines (Figure 1). In fact, an entire machine can be modeled either as a single unit or as a collection of components. If treating an entire machine as a single unit, the probability distribution of lifetime represents a summary of the combined effect of the lifetime distributions of each component.

If we have a model of an entire machine, we can jump to models of collections of machines. If instead we start with models of the lifetimes of individual components, then we must somehow combine those individual models into an overall model of the entire machine.

This is where the math can get hairy. Modeling always requires a wise balance between simplification, so that some results are possible, and complication, so that whatever results emerge are realistic. The usual trick is to assume that failures of the individual pieces of the system occur independently.

If we can assume failures occur independently, it is usually possible to model collections of machines. For instance, suppose a production line has four machines churning out the same product. Having a reliability model for a single machine (as in Figure 1) lets us predict, for instance, the chance that only three of the machines will still be working one week from now. Even here there can be a complication: the chance that a machine working today will still be working tomorrow often depends on how long it has been since its last failure. If the time between failures has an exponential distribution like the one in Figure 1, then it turns out that the time of the next failure doesn’t depend on how long it has been since the last failure. Unfortunately, many or even most systems do not have exponential distributions of uptime, so the complication remains.

Even worse, if we start with models of many individual component reliabilities, working our way up to predicting failure times for the entire complex machine may be nearly impossible if we try to work with all the relevant equations directly. In such cases, the only practical way to get results is to use another style of modeling: Monte Carlo simulation.

Monte Carlo simulation is a way to substitute computation for analysis when it is possible to create random scenarios of system operation. Using simulation to extrapolate machine reliability from component reliabilities works as follows.

1. Start with the cumulative distribution functions (Figure 1) or reliability functions (Figure 2) of each machine component.
2. Create a random sample from each component lifetime to get a set of sample failure times consistent with its reliability function.
3. Using the logic of how components are related to one another, compute the failure time of the entire machine.
4. Repeat steps 1-3 many times to see the full range of possible machine lifetimes.
5. Optionally, average the results of step 4 to summarize the machine lifetime with such metrics such as the MTBF or the chance that the machine will run more than 500 hours before failing.

Step 1 would be a bit complicated if we do not have a nice probability model for a component lifetime, e.g., something like the red line in Figure 1.

Step 2 can require some careful bookkeeping. As time moves forward in the simulation, some components will fail and be replaced while others will keep grinding on. Unless a component’s lifetime has an exponential distribution, its remaining lifetime will depend on how long the component has been in continual use. So this step must account for the phenomena of burn in or wear out.

Step 3 is different from the others in that it does require some background math, though of a simple type. If Machine A only works when both components 1 and 2 are working, then (assuming failure of one component does not influence failure of the other)

Probability [A works] = Probability [1 works] x Probability [2 works].

If instead Machine A works if either component 1 works or component 2 works or both work, then

Probability [A fails] = Probability [1 fails] x Probability [2 fails]

so Probability [A works] = 1 – Probability [A fails].

Step 4 can involve creation of thousands of scenarios to show the full range of random outcomes. Computation is fast and cheap.

Step 5 can vary depending on the user’s goals. Computing the MTBF is standard. Choose others to suit the problem. Besides the summary statistics provided by step 5, individual simulation runs can be plotted to build intuition about the random dynamics of machine uptime and downtime. Figure 3 shows an example for a single machine showing alternating cycles of uptime and downtime resulting in 85% uptime.

Figure 3 A sample scenario for a single machine

Benefits of Machine Reliability Models

In Figure 3, the machine is up and running 85% of the time. That may not be good enough. You may have some ideas about how to improve the machine’s reliability, e.g., maybe you can improve the reliability of component 3 by buying a newer, better version from a different supplier. How much would that help? That is hard to guess: component 3 may only one of several and perhaps not the weakest link, and how much the change pays off depends on how much better the new one would be. Maybe you should develop a specification for component 3 that you can then shop to potential suppliers, but how long does component 3 have to last to have a material impact on the machine’s MTBF?

This is where having a model pays off. Without a model, you’re relying on guesswork. With a model, you can turn speculation about what-if situations into accurate estimates. For instance, you could analyze how a 10% increase in MTBF for component 3 would translate into an improvement in MTBF for the entire machine.

As another example, suppose you have seven machines producing an important product. You calculate that you must dedicate six of the seven to fill a major order from your one big customer, leaving one machine to handle demand from a number of miscellaneous small customers and to serve as a spare. A reliability model for each machine could be used to estimate the probabilities of various contingencies: all seven machines work and life is good; six machines work so you can at least keep your key customer happy; only five machines work so you have to negotiate something with your key customer, etc.

In sum, probability models of machine or component failure can provide the basis for converting failure time data into smart business decisions.

Read more about  Maximize Machine Uptime with Probabilistic Modeling

Read more about   Probabilistic forecasting for intermittent demand

The Covid19 pandemic has placed unusual stress on public transit agencies. This stress forces agencies to look again at their spare parts planning processes, which is a key driver up ensuring uptime and balancing service parts inventory costs.

This blog focuses on bus systems and their practices for spare parts management and planning. However, there are lessons here for other types of public transit, including rail and light rail.

Back in 1995, the Transportation Research Board (TRB) of the National Research Council published a report that still has relevance. System-Specific Spare Bus Ratios: A Synthesis of Transit Practice stated

The purpose of this study was to document and examine the critical site-specific variables that affect the number of spare vehicles that bus systems need to maintain maximum service requirements. … Although transit managers generally acknowledged that right-sizing the fleet actually improves operations and lowers cost, many reported difficulties in achieving and consistently maintaining a 20 percent spare ratio as recommended by FTA… The respondents to the survey advocated that more emphasis be placed on developing improved and innovative bus maintenance techniques, which would assist them in minimizing downtime and improving vehicle availability, ultimately leading to reduced spare vehicles and labor and material costs.

Grossly simplified guidelines like “keep 20% spare buses” are easy to understand and measure but grossly mask more detailed tactics that can provide more tailored policies that better steward taxpayer dollars spent on spare parts while ensuring the highest levels of availability. If operational reliability can be improved for each bus, then fewer spares are needed.

One way to keep each bus up and running more often is to improve the management of inventories of spare parts – specifically by forecasting service parts usage and the required replenishment policies more accurately. Here is where modern supply chain management can make a significant contribution. The TRB noted this in their report:

Many agencies have been successful in limiting reliance on excess spare vehicles. Those transit officials agree that several factors and initiatives have led to their success and are critical to the success of any program [including] … Effective use of advanced technology to manage critical maintenance functions, including the orderly and timely replacement of parts… Failure to have available service parts and other components when they are needed will adversely affect any maintenance program.

As long as managers are cognizant of the issues and vigilant about what tools are available to them, the probability of buses [being] ‘out for no stock’ will greatly diminish.”

Effective spare parts inventory management requires a balance between “having enough” and “having too much.” What modern service parts planning software can do is make visible the tradeoff between these two goals so that transit managers can make fact-based decisions about spare parts inventories.

There are enough complications in finding the right balance to require moving beyond simple rules of thumb such as “keep ten days’ worth of demand on hand” or “reorder when you are down to five units in stock.” Factors that drive these decisions include both the average demand for a part, the volatility of that demand, the average replenishment lead time (which can be a problem when the part arrives by slow boat from Germany), the variability in lead time, and several cost factors: holding costs, ordering costs, and shortage costs (e.g., lost fares, loss of public goodwill).

Innovative supply chain analytics and spare parts planning software uses advanced probabilistic forecasting and stochastic optimization methods to manage these complexities and provide greater parts availability at lower cost. For instance, Minnesota’s Metro Transit documented a 4x increase in return on investment in the first six months of implementing a new system. To read more about how public transit agencies are exploiting innovative supply chain analytics, see:

• Optimizing Parts Management at BC Transit
• Smart Software and Metro Transit
• Smart Software Helps Metro-North Railroad Keep the Trains Running On-Time

I’ve stood in front of thousands of students. They’ve been more or less young, more or less technical, more or less experienced – and more or less interested.  I’ve done this as a university faculty member since 1972, first at Massachusetts Institute of Technology, then at Harvard University, finally in the School of Engineering at Rensselaer Polytechnic Institute. Between Harvard and RPI I dropped out of academia temporarily to co-found Smart Software with Charlie Smart and Nelson Hartunian. So since then, I’ve also been busy training business users to exploit the power of advanced analytics for forecasting and inventory optimization.

As I write this, I’ve just returned to my office at RPI after introducing first-year Industrial Engineering students to the basic concepts of inventory management. If they stick with the program, they will go on to take required courses in supply chain, system simulation, statistical analysis, and optimization. I told them stories about how useful they will be to their companies should they decide to make a career in the world of supply chain. If I’d had more time, I would have mentioned how capable they will be when they graduate relative to many of their corporate peers. These freshmen and ready and willing to stay the course, soaking up all the techniques and theories we can throw at them, and honing their practical skills in summer jobs or coop assignments.

What I didn’t tell them is that many of them will have to work to keep their intensity when they are on the job. It’s a sad truth that, for whatever reason, many inventory practitioners settle into a kind of stasis that impedes their companies’ ability to exploit the latest technologies, such as cloud-based advanced demand forecasting and inventory optimization. Gather enough of such people in one place and agility and improved efficiency go out the window.

I think one of the factors that dulls people is that the process of implementation frequently feels painfully incremental and prolonged. It often begins with a sobering inventory of relevant data, its correctness, and its currency. Then it moves to an often-awkward discovery that there really is no systematic process in place and the subsequent need to design a good one going forward. Next is the need to learn to use a new software suite. That step involves learning new vocabulary, some level of probabilistic thought, an ability to interpret new graphs and tables, not to mention a new software interface.  All this takes time and effort.

We’ve found that a few things help new customers stay the course. One is having a champion among management, an executive sponsor, who can vouch for the commercial importance of a successful implementation while ensuring the users are supported with continuing education.  A second is identifying and training a super-user or two having unusual combinations of technical and communication skills.  A third is breaking the training into bite-sized chunks and testing for comprehension after each chunk and repeating this process until it is clear that the new concepts, vocabulary, and process are fully absorbed. But all those maneuvers will come to naught without management being all-in and ready to stay the course.  Inventory planning practices in place for many years are not going to be replaced entirely over a three-month implementation process.  You’ve got to want it to get it.