If I made a survey of the Christmas wish lists for all supply chain executives ‘Better Customer Forecasts’ would be in the top 5 of almost every list, and number one on most of them. It’s one of the most popular topics in supply chain literature and for good reason.  Forecast too little and you won’t be able to supply your customer.  Worse, someone else might, and that might lead to permanent loss of business.  Forecast too much and you invest in resources that aren’t needed and may not be needed for a long time, if ever.  This is the most popular entrance ramp to the slippery sloped highway to financial oblivion.

 Much of what I see published on this topic pushes software services and systems that enable more ‘real time’ collaboration with customers and suppliers, aligning supply with demand, statistical analysis of customer performance, and the list goes on.  These systems can be very useful tools and potentially the missing link in achieving superior supply chain performance.  There are also many things that go into the successful conception, design and deployment of these kind of systems, I will not restate those here (Maybe in a later post).

Understanding your customer is the key concept that is so often misunderstood or neglected in the pursuit of better forecasting.  Almost every customer will send a bad forecast once in awhile. Most of the time the customer will catch them. It’s the ones where the customer doesn’t understand how far off they really are that can do damage.  You need to understand the difference between what they say they can do and what they really will do.  Don’t make the mistake of booking good news that can never happen.

The lean concept of ‘Takt time’ is very important here.  Most all systems have built in constraints that on one hand frustrate our ability to do more or go faster, but are very useful if we watch for it and use it to our advantage.  How many of you understand your customer’s constraints?  How much can they feasibly use in an hour, day or week? What is their theoretical maximum capacity?  Are they even measuring it?  How closely have they come to achieving this maximum capacity and for how long a time?  Running at 50% of theoretical is not uncommon.  How many days in a row have they been able to do 80% of their maximum?

Let me make an example.  A customer buys one type of widget from you and they forecast usage of 1,000 per week, but over the last year, they averaged 500 per week and have used 800 or more only twice.  A great deal of your material lead time is 12 weeks or more.  Your customer gives you a forecast for 1,000 units per week for the next three months.  If you booked this order and unless the constraint has been broken, they will use half of this amount and you will be sitting on a greater and greater pile of excess inventory.  Assuming no downstream intervention, the excess could grow on average by 500 units per week to a maximum of 6,000 units – equal to the average quarter consumption.  I’m not sure how you track inventory turns and performance, but this isn’t good by any measure.

The example above is obviously a worst case scenario.  Lead-time is the real enemy in this situation, but if you had a near zero lead-time supply chains, you wouldn’t be reading this. How this would work out in the real world would depend on what kind of material commitment horizon arrangements exist between the various customers and suppliers along the chain.  I know it is very dangerous to load numbers different that what your customer gives you, but you better speak up if you think they are sending out instructions that don’t make sense.  Knowing the system better than they do is how you increase your value to them which is always a good thing.  If the customer doesn’t listen – and some don’t, well, find ones that do.

As I said before, there are a lot of good software tools that will help you manage your supply chain.  Unless you understand what is behind the numbers going in and the limitations of what comes out, you can make just as many mistakes and even more quickly than before. Know what you can really do and expect to do.  What your customer can do and what your suppliers can do. If you know this, you are well on your way to putting more advanced supply chain software to use.  If you don’t, you won’t get your return on your system investment.

By the way, this concept is even more important in the reverse situation. Change the word ‘customer’ with ‘supplier’. Do you know your suppliers constraints?  Do they know? Have you buffered your production peaks if they can exceed your suppliers?


Leave a Reply