If you are reading this, we are sure you agree that demand forecasting is really important! But why is it so important for you and your business?
Our opinion is that you would like to predict demand to avoid two negative effects:
- Going out of stock and having no item to satisfy demand
- Ending up with a large stock of unsold items
The first effect is really bad: you miss opportunities, your customers will be dissatisfied and maybe they will never come back to you again. Not only you lose some sales: more important, you are losing customers and reputation. With social media around, an unsatisfied customer will give you a negative evaluation and convince many others not to trust you! It is very difficult to build a good reputation. It is so easy to lose it. So out-of-stock is the absolute evil!
But consider the second negative effect: large inventories. Your customers will be really happy with all that availability. But will you be happy too? I guess you should not: large inventories mean many things, most of which negative…
You will have direct costs (for storage, insurance, refrigeration, to name a few), but also many indirect, hidden costs: capital immobilization (you might have used your money for something else), obsolescence, seasonality (you won’t need to keep that many ice creams when outside it’s freezing).
So in our opinion you don’t need “just” forecasting: you desperately need inventory control and optimization! In this note we will introduce the basics of inventory forecasting.
The Ideal: Perfect Forecasts
Maybe your business is in the happy situation in which your future sales have no uncertainty at all. For example you are producing in a make-to-order situation and you perfectly know, for a few weeks or months in advance, your future demand.
Very easy! Or not?
Even with perfectly known demand, you should plan production and buy raw material or products. When, what, how much are you going to buy and produce? Of course you will always be ready to satisfy your demand, but your decisions will influence your costs: are you ready to profit of discounts on large lots buying? Or to take advantage of marginally decreasing production costs when your production lot is larger? Or can you efficiently deal with transaction costs when ordering, or with setup and changeover costs in production?
These facts lead to larger quantities, but keep in mind that this means inventory! Even with a perfect forecast your decision will not
be trivial, and chosing a suboptimal policy might mean losing a lot of money.
An optimal balance is needed to exploit the benefits of large lots without building too costly inventory.
The Usual: Uncertain Future
Things get worse, of course, when facing an uncertain market, in a make-to-stock situation in which you should build inventory, but you don’t kniw exactly when, what and how much your customers will ask in the next period.
Uncertainty, per se, is not the problem: the problem is not taking it in due consideration. There are many factors which, if not taken into account, will lead you to take bad decisions. Here is a list of some:
- Trends and seasonality in your past observed demand
- Product life cycle
- Product substitution and cannibalization
All these are related to demand forecasting: when there is uncertainty you should take the most from your past data. But there are other important factors to be carefully considered:
- Variability: assume your forecasted demand is 100. It makes a big big difference if this is because in the past your demand has been either 99 or 101, or if many times you had no demand and about as many you had 200. You should take the variability around your expected demand into careful consideration.
- Lead times: your inventory is quickly going to zero: you are happy, your business is going well! But you should keep it going this way or even better, so you quickly place an order. When will your order be delivered or produced? The time from your order to its delivery is called lead time. It might be really short, but often it is days or weeks. It may also be uncertain. If you underestimate it, you will run out of stock and lose demand and customers.
- Costs: of course you are concerned with costs, it’s your business. But have you considered that, taking costs into consideration, even if your demand forecast is 100, it might not always be a good idea to have a stock of 100 or more?
Sometimes you will prefer to have 90! Sometimes 110… Why and when? If your product’s life cycle is short (a fashion item, a new ice cream, fresh food, newspapers, ….) you will have to take into account your production or your buying costs, your revenues for the demand you can satisfy, your costs for final inventory: this might be the lower revenues when selling at discount or the additional costs when you will be required to manage inventory disposal.
The ratio between your expected gain on satisfied demand and the expected loss when you end up with an inventory might sometimes guide your best decision quite far from your forecast.