You are currently browsing the daily archive for March 23rd, 2007.
This example shows how reducing the response time of the supply chain reduces the variation in supply and provides better control.
The example product was made in a single factory and shipped worldwide through a number of distributors to supply a market with a steady daily demand. In other words, the true demand was completely flat. The stock and ordering policy of the distributors adds variation which is reflected in variation of the factory’s output. At the start of the period shown in Fig 3, the manufacturing response time was 18 weeks which coincides with the time between peaks of supply. It can also be seen that the supply quantity varied by +/- 30% representing quite a strain on factory management (feast and famine). Batches of the same product were made every 6 weeks and manufacturing lead time was 12 weeks for some of the components.
Fig 3
During the middle of this period, the factory was changed so that lead time for the most-awaited component was reduced to 8 weeks and batches made for shipment every 3 weeks. The right part of the graph shows that the peak to peak time reduced to 11 weeks (the new response time) and the amount of variation reduced to +/- 15%. In this new regime, everyone was much happier and felt more in control. A fringe benefit (actually probably ultimately worth more in Sales) was that the customers received fresher product and could keep it for longer before it reached the end of its shelf-life.
A further observation is that the reduced variation did not happen as part of a conscious effort on anybody’s part. It seems that the distributors “sensed” the greater responsiveness of the factory and adjusted their ordering and stock policy unconsciously. In other words, it is human psychology that puts the variation into the demand (“I don’t trust the factory to deliver when they say, so I will order a bit more, in case”. Followed a few months later by “I seem to have far too much, I’d better cut back drastically on my orders”). Greater responsiveness leads to greater trust and a more realistic ordering pattern.
Fig 4
It is worth also mentioning that manufacturing cost was held neutral, whilst inventory came down. The chart shows on the same timescale how the inventory level of 3 of the components reduced as implementation progressed.
by James La Trobe-Bateman, reMODEL Consultants International Ltd
In most industries, demand varies in the short, medium and long term. Such variability is why demand forecasts are inaccurate. Fluctuations in demand are notoriously difficult to predict and this has driven a trend towards shorter lead-times and “make-to-order” production. If demand in a factory was stable, its output could be set to guarantee satisfying that demand. Even when demand changes, such a factory can respond within its capacity if the forecasts are accurate. In both situations, the time it takes to process material or the interval between making batches would have no effect on the success of the factory in meeting demand.
However, in reality, demand does change, new opportunities for sales present themselves and existing customers change their minds about product mix. These situations are impossible to forecast accurately. So, in real life, the factory must respond to those demand changes. At those moments of unexpected changes in demand (and only then), it becomes apparent whether the factory responds like a supertanker or a jet-ski. Various aspects of product and process design influence responsiveness, but their effect can be reduced to two measurements: the interval in time between batches and the time it takes to process a batch. “Response Time” is a measure of the ease with which manufacturing can change its output to meet changes in demand.
Definition
Response Time = Batch Interval + Process Time
An Analogy
It is helpful to think of it by analogy. If you were planning to fly from A to B, how long should you allow to be sure of getting there?
Fig 1
i.e. Allowed Time = Interval between flights + Flight Time
which equals 1 + 2 = 3 hours in this example.
If you arrive just in time to catch a flight, it will take you 2 hours, but if you just miss one, it will take 3. In the absence of a flight timetable, you will need to allow 3 hours to be sure.
It is possible to talk about the response time for intermediate components too. However, it only makes sense if the intermediate is stocked: “planes must land at the airport”
Ways of Reducing It
It is clear from the definition that response time is reduced by:
• Making batches more frequently (reducing batch size)
• Reducing process lead time
However, the most common form of response time reduction is by the use of stocking points, in particular, finished goods. Sales can then be in units (=batch size) of 1 and processing time is typically just the shipping time from distributor to customer.
Of course, buffering against sudden changes in demand by adding raw material, intermediate or finished goods reduces response time at the expense of inventory. So response time reduction should not be viewed as an end in itself. It needs to be done without increasing inventory or cost to be a credible improvement to the business.
Fig 2
A response time can be calculated for each set of processes which end in a buffer stock. The complete supply chain is thus an assembly of such links each with its own response time.
Thus there is generally more than one figure to be looked at, although the final one (nearest to the end customer) is the one most often scrutinized.
The Result
A more responsive factory:
• Is less likely to default on customer orders
• Has fewer disruptive changes to production plans
• Is in less of a panic
Fit with Management Strategies
Whilst not stated explicitly, Response Time reduction is the goal of many manufacturing strategies such as “Lean Manufacturing”. However, there is nothing new about it as a concept. Henry Ford can be thought of as an early advocate when he established a car production system that smelted 1500 tons of steel a day to turn out a car every 49 seconds in the shortest possible elapsed time.
It is a common sense concept. If you want to satisfy the customer, you must respond to their demands.
By James La Trobe-Bateman, reMODEL Consultants International Ltd
see Case Study for an Illustration of the Real Effect of Reducing Response Time
