How time buckets limit ERP’s ability to do resource planning
Mike Liddell, CEO of Lean Scheduling International and Author of “The Little Blue Book On Scheduling”
It is a human trait to make sense out of time events by putting them into time buckets such as days, weeks, months and years. Knowing this makes it difficult to accept that time buckets should never be used as a mechanism for scheduling.
Unfortunately, the one thing that ERP systems have in common is they all employ the same outdated resource planning techniques that were used back in the early 1980´s such as infinite capacity, backward scheduling and time buckets.
Whereas much has been written about the limitations of infinite capacity planning, very little has been written about the crippling problems time buckets cause when they are used to do detailed scheduling. This article pulls the covers off and exposes key problems that can be clearly blamed on the way that ERP planning uses time buckets.
Understanding Cause and Effect
The whole purpose of any scheduling activity is to help the scheduler understand the likely consequences of a decision or a change to the schedule. We call this process understanding cause and effect. The underlying problem is the way ERP manages time which is rooted in the level of detail needed.
For example, if I wanted to book a meeting room from 1 pm to 2 pm on Tuesday afternoon with two of my managers, I would need a scheduling system that handles precise time. If I have a scheduling system that uses daily time buckets like ERP does, then all I need to do is consume 1 hour of the meeting room´s capacity on that day. Unfortunately, by using this process, all anybody would know is that the meeting was planned to take one hour on that day. With no exact time, managers would be waiting around until the meeting started. Also, if someone else tried to book the meeting room for that day then the problem would multiply and the chances of causing confusion would quickly increase.
In a typical manufacturing plant, the likelihood of confusion is much worse because this happens thousands of times a day. Most events are also dependent on other events (for example you can´t do step 2 until you have completed step 1).
Using time buckets in the manufacturing world has other, even more serious, limitations. For example, you could have 5 or 6 jobs lined up on a specific resource on a specific day. Because time bucket logic doesn´t let you create a sequence for those jobs, all it can do is assume that each of these events will be finished by the end of the day. In the real world however, one event could be finished and ready to perform its next operation by 8:30 that morning. But this possibility cannot be handled using time buckets.
Power of Sequencing
Every scheduler understands the power of sequencing. They know that the sequence of jobs at a resource can impact what happens to every other job in the plant. Because time bucket logic doesn´t allow you to determine how jobs are sequenced, the scheduler is unable to predict the likely consequences of his decisions. And it gets worse! Without knowing the job sequence, it is not possible to calculate things like sequence dependent setup times or projected queue times. In order to get around this limitation, time bucket systems must assume average setup times and average queue times which are almost always wrong.
Not being able to manage cause and effect is a somewhat unfortunate limitation for a system that calls itself an Enterprise Requirements Planning system. Planning with ERP is like driving a car with your eyes closed. The only time you are likely to know you have a problem is when you hit something and by then it is usually too late to solve the problem. A more detailed explanation of these issues can be found in my book titled “The Little Blue Book on Scheduling”.
ERP systems have been around for some time because they do so much more than planning. Make-to-stock manufacturers have had limited success with ERP planning because they have built inventory buffers that cover the inherent inaccuracies. As manufacturers move towards leaner, make-to-order business models, they are starting to realize they need a better way to plan and schedule.
The good news for these manufacturers is that all the problems identified in this article can be resolved by adding a good APS (Advanced Planning and Scheduling) system to their existing ERP system. The major differences between an APS and ERP system is APS uses a “scheduling engine” which is capable of accurately handling time at the detailed level and at high speed.