Reducing Fluctuations Downstream

This last post of my series looks at fluctuations that originate downstream from your location. In other words, how to reduce fluctuations originating from your customer. Granted, this often is the most difficult one, as you usually have not so much influence over your customers (unless you have a monopoly). Let’s have a look.


Source Make Deliver FluctuationsAgain following the source-make-deliver structure, we now look at “deliver.” This is your customer. It may be the end customer; however, it could also be simply the next stage in the value stream, which again is your customer.

As always, there is the possibility to decouple using buffer inventories. Or you could adjust the capacity to better match the customer demand. However, these two approaches address only the symptoms and not the cause of the fluctuations.


Pizza Delivery

Your overall goal in reducing fluctuations originating from your customer is to have your customer order more regularly. Unfortunately, you usually don’t have enough influence in simply telling the customer when to order. However, even then there are some methods that nudge your customer to order more regularly.

Sales Promotion

One popular method is a sales promotion. You reduce the price to increase sales. Similarly, you could also increase the price to reduce sales. By adjusting the price, you can change the customer behavior. However, this also affects profit. Maybe you sell more items but with lower profit, or less items but with higher profit, in any case it has an influence on the profit. If it helps you to make more profit, then you should have done this already just for the money. If it reduces profit, you have to be careful that the cure of lower profit is not worse than the disease of less fluctuations.

A similar approach that keeps the price stable is a marketing campaign. You do not change the price, but merely motivate more people to buy your stuff. But be aware that marketing is a bit of a magic black box that does not always work.

These approaches have some delay between the implementation and the change in the behavior. Hence, it is more for longer-term fluctuations. It is also helpful if you can predict a change in demand beforehand. A key example is seasonality, but it may also be an upcoming legal change, or an increase or decrease in the sales tax. In any case, by carefully adjusting these levers, you may be able to reduce the magnitude of fluctuations.

This does require coordination between sales on one side and production and/or logistics on the other side. I was on the receiving end of a very successful sales and promotion campaign that hit my shop floor with increased orders when we were not even able to satisfy the initial demand. It increased the fluctuation and led to many delayed deliveries, but sales looked good because they sold stuff. Anyway, poor coordination can lead this to backfire.

Expand Markets

Another example to counteract fluctuations is to expand markets. This can work in multiple ways. In general, the larger the sales volume, the smaller the fluctuations compared to the sales volume. Hence, if you sell more, your fluctuations grow slower than the output. And, management usually does not mind if the profit grows along with the increased sales.

Your anti-cyclic customer
Your anti-cyclic customer

If it is a seasonal fluctuation that you want to reduce, you should try to expand into new markets that have an anti-cyclic behavior. Skiing gear usually is highly seasonal, but in Australia the skiing season is July and August. Unfortunately, there are only 25 million people in Australia, plus another 5 millioon in New Zealand. Don’t you wish Australia and New Zealand would have more people?

New Product Variants

Nordic Walking
Your anti-cyclic product

There may also be a possibility to introduce new products that have an anti-cyclic seasonal behavior. If you are making ski sticks, consider expanding into hiking sticks. The whole idea of Nordic Walking originated from the desire to have a product that you can sell in summer.

Reduce Product Variants

Sales and Manufacturing DreamAnother option is to reduce product variants. This may be confusing, since the previous suggestion was to add new product variants. However, the additional variants were for anticyclic markets. The reduced variants are for the “normal” market.

This works in two ways. First, reducing product variants will (hopefully) increase sales of the reduced variants. Higher quantities mean automatically a more stable demand.

One of the many variants for glove compartments…

Second, it means less components and products to track. You also get higher demands for fewer components. A common example is cars. If you buy a Toyota, there are a few model variants, with few options. Most Western car makers have many more options, which increase the fluctuations and subsequently the overall cost. An Audi A6 has over 150 variants for the glove compartment alone, and 18,800 variants for the door cover. Think of the logistic nightmare of having 150 different glove compartments for the same car! Reducing some of these variants will have a lot of benefit, one of which is the reduction of fluctuations.

Be aware that you may face a lot of resistance from sales and management in general when you want to reduce product variants. Their argument is that having more different products allows you to capture more market niches. They are, of course, also correct. Normally you would do a cost-benefit analysis, but in this case there is probably no hard data neither for the extra market niche nor for the reduction in fluctuation. Such situations where you have to argue based on opinions rather than facts can quickly get messy.

So, as you can see, there are some options for reducing fluctuations on the customer side. Of course, these require effort and may also cost some money. You have to decide which one is worth the cost. Also, I never said it is easy. Overall, it may be easier and better to start reducing fluctuations with “source” or “make” rather than “deliver.” Nevertheless, I hope this gave you some inspiration. If I forgot anything that you think can also help, please let me know. Now, go out, reduce your fluctuations, streamline your system, and organize your industry!

Series Overview

4 thoughts on “Reducing Fluctuations Downstream”

  1. I liked your summary of the challenge on dealing with sales/marketing when pushing to reduce variants (Say glove boxes): Marketing will claim (sans data) the additional variant will add a handful of new sales per year, but the value of each car is quite high, so there is a big revenue carrot….while the manufacturing cost of managing complexity always seems to not reflect the true ‘cumulative’ cost. One more glovebox, when you already have kiting/error proofing, pick to light etc carries a small additional manufacturing/quality cost….yet emotionally, the cumulative impact of massive variety is, instinctively, huge. One of those cases where we may know the price of complexity per component, but perhaps not the true value to the enterprise of the complexity.

  2. Beside marketing activities you may also reduce “hidden waste”, so called by professor Bicheno, University of Buckingham: Reduce internal fallout, because this will lead to reordering and add to customer fluctuation. Also reduce lead times to be able to react faster to customer fluctuation and finally add proper buffers in time, capacity and inventory to reduce the effects of customer fluctuations.
    I also would like to add that often you have “internal customers”, for example in plant to plant deliveries. This is a good starting point for companies to analyse customer fluctuation and to decide in which value stream to start with improvement activities,


  3. A good and not often discussed topic, I feel. Level selling, hand-in-hand with heijunka, should provide the required stability to your operations. But heijunka is just a compromise between your operations and an unlevel demand. I fully agree it should get more attention.

    In addition to your list of possible actions, others that could help are, for instance:
    – segmenting demand patterns (stable vs unstable) and separating (allocating) these internally, which at least isolates the variation;
    – reducing failure demand which may create unwanted and unnecessary fluctuations;
    – timetabling and order slotting, which is to play with lead times to level out the pattern.

    In the end, though, the thing I mostly wonder about is that as a supplier, companies often desire regular, leveled demand from their customers, but as a customer, they are not ordering similarly with their supplier. This is strange, as they know that their desire on the other side of the building comes from the gains they can have from such a pattern. So, as we are all both customer and supplier, why can’t we seem to create a win-win?
    I wrote about this leveling paradox in an earlier blog post here if interested:

  4. Hi Christoph,
    your article is very interesting as always!
    One thought: reducing the arrival variability is certainly very important, but it can be done up to a certain level, after which it is necessary to buffer the variability with stock, capacity or lead time.
    I think also that if reducing variability compresses revenue more than it reduces the cost of buffers, it’s not worth doing.

Leave a Comment

Cookie Consent with Real Cookie Banner