We know better than to try to force fit strategies yet time and again we find businesses planning and managing their processes with a ‘one size fits all’ approach. Hospital services, insurance underwriting, pharmaceutical production, consumer electronics assembly, integrated supply chain services all struggle with finding the right balance of specialization and common process. At the extremes there is the chaos of infinite flexibility where every order, customer, patient is an exception to the other extreme where the ‘cats and dogs’ are blocking traffic on the highway.
Simply sorting process demand (orders, service requests, projects, patients) by volume or frequency isn’t good enough.
Adding the dimension of demand variation is an improvement, but optimizing inventory, process performance, supply chain doesn’t end with a 2×2 graph of Volume vs. Coefficient of Variation. But before we tackle total cost, margin analysis, supply chain policies let’s go over the fundamentals.
Gather up representative data for orders, arrivals, requests – whatever you have that represents demand on your process, business, factory, enterprise. In HR this might be recruiting requests, in product development an inventory of active projects, in a factory grab customer orders. [ABC Analysis: how to]
Now take the history and summarize product, or service and make a PQ Product vs. Quantity Pareto histogram like this …
Sort the product or services from high to low, left to right. The highest volume activities want to be streamlined, maybe even have dedicated processes and resources. Call these A items the ‘highway’. The lowest volume of orders or frequency of service items, the C items may need to be handled on a case-by-case basis. If you don’t then you’ll have a ‘mixed model’ process where the A’s and C’s get in each others way.
–A’s = dedicated cell or line or department
–B’s = mixed model
–C’s = “craft” or specialty work cell, or premium service department
ABC PQ Analysis only goes so far and as previously discussed [1
] doesn’t consider that while most high volume items are usually stable and predictable some of the high volume items may be unpredictable and have erratic demand, and many of the low volume items will be well-behaved and very repeatable.
If you made a time series graph for every product or service you would be able to ‘eye ball’ the period to period variation in demand and perhaps make some judgements on categorizing similar profiles.
Here are three examples from a recent pharmaceutical demand analysis.
The first product has fairly stable weekly average demand. A weekly Coefficient of Variation
of 0.33 might be considered a little high by some. The spike at around week 52 is a year-end effect, so overall this product is well-behaved.
The second product is half the volume of the first visually more ‘noisy’. A weekly Cv indicates that this product would be a good candidate for pack-to-order, otherwise significant inventory would be need to be able to cover the spikes.
Product 3 is very unstable. Any guess on how much inventory to carry, raw material to buy, capacity to reserve, or lead time to promise is certain to be wrong. This product would be a good candidate for make-to-order, move to a low volume – high mix business unit (and price and staff accordingly), or possibly discontinue.
Any attempt to try to efficiently run these three products through the same forecasting, planning, scheduling, production, distribution process is bound to be flawed. The lead times, lot sizes, equipment assignment, and other planning parameters should all consider the item’s demand pattern. Where you find high Cv ‘cats and dogs’ items coexisting with low Cv ‘steady eddy’ products you need to seriously consider applying demand segmentation and not trying to make one size fit all.
Products and service have a lifecycle; usually new products volume increases and the initial demand is spotty but stabilizes as the product matures. Then new products compete and demand declines and becomes irregular for the first generation.
Customer demand can be described, quantified. Here’s how …
1. Pick product, product family, or business unit of interest
2. Determine an appropriate time unit – hourly, daily, weekly, monthly
3. Gather the true customer demand as best you can. Be careful about using promise dates instead of requested dates, and be doubly cautious of schedules which are often smoothed, filtered, or otherwise manipulated.
4. Create the graph or time series plot
5. Then calculate some simple descriptive statistics and interpret
Then use Excel Data/Data Analysis/Descriptive Statistics and interpret the results. In this example mean, median, mode are all close to 17 per day.
The range is 49 (max-min) and +/- 3 standard deviations is 34, or -16 to 50 (17+/-3*11)
(or zero to 50 since we don’t usually have negative customer orders, and we don’t usually worry if demand is lower than the mean, it’s when demand is far greater than the mean that things get interesting.)
The standard deviation is less than the mean, so the demand pattern is somewhat stable.
This demand is a bit skewed to the right, but only slightly;
confidence limit estimate =2*SQRT(6/N) = 0.94
A confidence limit estimate for kurtosis is 2*SQRT(24/N) = 1.89.
Since 1.45 is within +/- 1.89, we can say that the data is marginally normal.
The confidence limit is smaller than the mean, so the mean is reliable.
Process Model Selection
- Margin Contribution & Cost to Serve may be more insightful than just looking at demand quantities or sales dollars.
- Products and services may have different demand, inventory, customer replenishment, supplier replenishment, sourcing and total landed cost, and other policies.
- Managing these policies in a dynamic business can be a challenge; so look toward optimization automation to stay ahead having the right strategies for the right products instead of one size fits none.