How to avoid the next “GFC” (Lean)

Firstly apologies to those who thought this post would be about the Geelong Football Club.

This post is about the other (less important) GFC otherwise known as the “Great Financial Crisis”.

In the last two posts we investigated forecasting. There are three components that can contribute to an accurate forecast, Trend, Deviation from the Mean, and Seasonality.

Should the item being forecasted display strong attributes in these areas then the forecast is more likely to be accurate, if it shows none of these then the forecast will be less accurate (the more random the history the harder to forecast).

Using petrol prices as an example.

Prices are slowly going up as we creep out of the economic slump (trend), its higher on the weekend than it is during the week (seasonality) and while we don't know what the exact price is going to be next week its highly unlikely to be less than a $1 or more than $1.60 (deviation from the mean).

However what happens when an unexpected (and catastrophic) event occurs?

Let’s use the stock market as a representation of the general economy and examine the affect of the collapse of the financials system (an unexpected and catastrophic event) on forecasts and the flow on effect of getting the forecast wrong.

All ordinaries from April 2003 until October 2007.

If you were to forecast based on this ‘history’ then your forecast for the following years would show strongly trending increase in demand. You would "plan" around this forecast.

And that’s exactly what manufacturers, distributors and retailers around the world were doing, ramping up production to meet this forecasted increase in demand.

However along came the crash (October 2007 to March 2008).

Our forecast based on 'history' 2003 to 2007 is now hopelessly inaccurate.

The effect on the economy is striking.

Manufacturers and distributors will have production schedules and inventory levels (and loans) based on their forecast of strongly trending increased demand. They now have too much stock, too many staff, be too highly geared and subsequently have lots of problems!

They will need to stop/slow production until this excess inventory is sold which leads to layoffs which in turns leads to reduced demand which is now feeding into a now strongly downwardly heading trend.

But how do we avoid this overshoot when the trend changes and avoid (or at least minimise) it’s nasty after affects?

Essentially it’s by only making and stocking items that you have already sold by not forecasting!

The challenge with this approach is that items with long lead times (i.e. the time it takes to get from raw material to finished good to deliver to customer) need to be forecasted.

Your customers want what they ordered now, not in six months time.

A solution to reducing these lead times is to enable a process where you are making and stocking only items you have either already sold or are likely to sell in the near future (i.e. only selling what customers actually want) rather than speculating on the mid to long term future.

This solution is generically called "Lean" Supply and Manufacturing, which we will discuss in detail in future posts.