Reducing forecast error and avoiding the results of inaccuracy

Reducing forecast error, reducing the need to forecast and avoiding the problems caused by inaccurate forecasts

Part 1. Why forecast, why forecasts are wrong, and why does that matter?

This article is part 1 of two articles on forecasting. This first article deals with the need to forecast, why forecasts are often wrong (contain errors), and the implications of forecast error. The second article (available on request) deals with improving the quality of forecasts, reducing the need to forecast, and reducing the impact of forecast error. This service is not available to consultants.

Links to More Best Practices and Training Below

Why do you need to forecast?

There are a number of reasons why you may need to forecast. Below are listed the ones we have encountered, together with their characteristics.

1. Product life cycle

The following diagram (figure 1.) illustrates a product life cycle. It shows a growth in volume up to a peak. This peak which may be short lived is sometimes called the novelty curve then declines to a plateau of relatively stable demand and then a decline into obscurity. The scale of the curves and slope of the lines varies from product to product but the general shape is applicable to many if not most products.

Figure 1. Product life cycle

The volume and duration where the changes in slope are most significant are:

The most important to forecast
The most difficult to forecast
The ones where people want to exert maximum influence

2. The Business Cycle

Together with economic prosperity and decline goes a cycle of what has come to be known as "the business cycle" (a period of slow or negative growth followed by the next period of rapid growth). It is also referred to as "boom and bust".

The impact of this on capacity and stock is shown in figure 2 below:

Figure 2. The business cycle

Again it is important to identify the next upturn or downturn accurately, or there will be the sort of implications shown in the diagram.

3. Supply chain lead-times greater than consumer required lead-time

Where there is uncertain demand and the supply chain lead-time is greater than the customer required lead-time, safety stock needs to be held in anticipation of an order. I.e. Items cannot be made to order.

4. To provide input into long term decisions such as capital investment.

Often major capital equipment or expansion plans have to be viewed against a forecast of what the demand will be.

5. Peak demand and average capacity

This is the problem caused when customers want things at a faster rate than your capacity allows. For a period of time the mismatch ((anticipated peak rate of demand - peak rate of supply) x lead-time) has to be made in advance to ensure that sales are made when required. It is not readily realised that in order to give consistently high service levels, peak demand has to be matched by peak capacity. This gives accountants concerns since the capacity is therefore under-utilised for the remaining time. Most management accounting systems and budgetary control systems are based on top down spreading of annual targets which has a smoothing (averaging) effect, rather than bottom up costing of timed causal relationships, for example, a sales campaign leading to increased costs at a particular time. This would not be a problem if these budgets were not then used indiscriminately to control expenditure rather than intelligently to explain necessary variances. This cost tension then starts to create the problems shown in figure 2. Extreme examples of this problem are in seasonal demand such as vehicle batteries, and garden equipment and success stories where products are selling better than expected but cost pressures prevent increasing capacity. Much worse however is where products are going rapidly into steep decline but backward looking management accounting systems do not exert cost pressures until too late.

6. Spares

Spares present two major problems:

In early life the supply chain has to be primed on estimates of mean time between failures based on small sample testing.
In later life as the original equipment is taken out of production it is often necessary to estimate all time spares requirements with little knowledge of the decline in usage of the product.

Mid life spares forecasting whilst still potentially problematic for long life components, is less troublesome, since after the novelty phase (see figure 1) demand is more stable.

7. Transport costs

Transport costs inhibit the supply to order situation, which is the ideal lean supply chain (See Designing Lean Supply Chains), forcing suppliers to "batch up" to reduce unit transport costs.

8. Manufacturing costs

The old adage "cheaper by the dozen" is often used to justify large batches (See Participative Sales and Operations Planning), which then require a forecast to justify the stock holding costs and reduce the risk of obsolescence.

Why are forecasts wrong?

Which of the lines in figure 3 represents the best forecast after plot 3? There is a good argument for each.

Figure 3: Best-Fit Trend

There are a number of reasons for forecast error. These include:

Customers often make secret plans, or plans you are not party to, or no plans at all.
Often we use an imperfect source to use in our forecasts when better sources are available or we use the best source but do not receive the best information available from that source.
Information lags can lead to using old information where newer information is available, so that the new trend is spotted later.
Decision making and communication of decisions being omitted or delayed, e.g. a sales campaign.
Over reliance on one source without confirmation from other sources can give rise to local distortions of the information.
Conversely if too many inaccurate sources are used they may incorrectly weight the resulting forecast.
Demand distortion, where the information is correct but it has been subjected to a local distortion. An example I often use is the one where I went to three consecutive stores trying to obtain an item and they put it on order in all three. At the fourth store I obtained the item from stock. But I did wonder afterwards if the factory were making the other three.
Communications links may not have been established or are not established reliably, such that the resulting forecast is based on incomplete or partially out of date information. Changes of plan or environmental changes can of course happen. If good communications links are established these will be picked up.
Changes in purchasers' behaviour can occur for a number of reasons, but the ones, which should not be a surprise, are technology changes, dissatisfaction with your product or service or changes in competitors' behaviour (which you should know or at least anticipate).
There are a number of sales malpractices that cause demand distortion, such as 3 for 2 offers, January sales, etc. (See Managing Demand)
There are a number of inventory planning malpractices which can distort demand such as excessive batch sizes, or safety stocks. MRP1 systems in particular give rise to a phenomenon called "nervousness" which is caused by changes in planning parameters such as Bills of Material, safety stocks, batch sizes, lead-times, planing horizons, and scrap allowances. Also Bills of Material (See Best Practice of the Week 14 "Effective Bill of Material Design"), engineering changes (See Best Practice of the Week 022 "Change Control"), stock errors or PI stock corrections (See Best Practice of the Week 029 "Bin Discipline") add to this problem. With MRP2 and APS systems the problem is made worse by any change to routings, capacity, or work centres, priority rules or correction of inaccurate information.
Previous plan failures (or operational malpractices) can cause a surge in output as the problem finally gets fixed. This, when viewed in the future (in the next forecast), can be viewed as increased sales.
One of the great difficulties in this computer world is our over-reliance on computer models for forecasting. This spawns the following types of problem:
- Inaccurate models (Not recognising demand patterns such as seasonality or trend, or most importantly sparse or unusual demand on which mathematical forecasting models are unsound).
- Over complex models (because you can).
- Calculating safety stocks for all items in the same way. (A fast moving inexpensive item with many customers is quite a different risk to a slow moving expensive item with few customers, although they may exhibit equal standard deviation of demand).
- Unsubstantiated belief in an invalid model.
- Unsubstantiated belief in the data source(s), maybe simply because it is (they are) the most readily available.
- Precision verses accuracy. Computer systems are capable of doing calculations to many decimal places based on entirely inaccurate information. Be aware that you may be precisely wrong.

Not recognising the business cycle or product life cycle shown above.
Incomplete, due to the difficulty of collecting the data on a timely or reliable basis.
Inconsistent, due to not collecting the same data consistently in consecutive forecasts, due to some systems failure. (An apparent trend in this case could simply be simply due to some additional or omitted data.
The point at which demand is measured. If demand is measured at a point in the supply chain where there is distortion due to local factors, the resulting forecast will be wrong. For example if demand is measured by despatches and despatches are batched to provide full loads the demand will appear lumpy. It will also be distorted by failure to deliver orders on time (back orders) at this point. If you measure customer orders received, you are ignoring the underlying demand for orders you did not win for other reasons. If you measure enquiries you will be including artificial demand based on price checking exercises from customers. If you measure output you will be measuring production batches not demand. If you measure demand on your customers you will include their distortions and any current or previous sourcing decisions they have made, not their future ones.
And last but least natural variation which you can do nothing about, but is by far the least problem.

The implications of forecast error

The obvious implication of forecast error is an over or under reaction to the latest trend. This gives rise to risks of: