Tuesday, August 18, 2009

Forecasting as a ritual!!

Most companies struggle with forecasting process as it is indeed difficult to predict the future in a realistic way. Recently I was reviewing the forecasts and found that there are four major categories under which units can fall. I have listed them but it maybe that most companies have the same patterns.
1) Average demands prolonged over this year remaining and next year-no change month on month just same nos. Amazing as it may sound but this is what I call "pulling the excel cell to fill all the blanks". There is no value to such information (for example suppliers will just ignore such information as they know this is wrong from day 1).But still people fill the files and ritually send this back.
2) Erratic demands with a 200 k in one month and next month 9k followed by 350 k in the month after. Maybe we got supplies like this in previous years but to use this to forecast is disaster. We cannot have such abnormal variations based on historical supply records as it will also turn out wrong. Puristic forecasting only on historical patterns which are skewed is a disaster from day 1.
3) Conservative forecasting with a low 2010 and a high coming months. Maybe the focus is on this year but we cannot have this patterned like this. For example growth economies have a forecast fitting into this pattern with some cases supplier getting much lower demands progressively (in some cases moving to zero with no alternative suppliers)..why do people do this..just to ensure they don't pre-empt business plans. Sand bagging in every step. Recipe for unadulterated diaster.
4) Real forecasts-some units do a fantastic job of forecasting supplies with a clear move of suppliers (share/volumes etc) with clear timings.
I remember a story:
Once a king provided a 1 liter dark coloured jar to each of the houses in the kingdom and said he wanted each house to pour in a concrete well one liter of milk on a new moon night. The milk will be used for feeding orphan children next day. There was a big queue by 10 pm on the appointed night and each one poured their share in the well. Most thought that since it is dark and since most others will pour milk he pouring one liter of water will not be noticed. So the smart ones choose water instead of milk. Next day morning the King showed the well which had 99% water and 1% milk. He said" The milk poured by the honest people are also wasted due to the water pourers". The country learnt the lesson. I am sure we are in the same situation as people who have done a good job in forecasting will not get what they want due to others who choose to do a linear progression or just give the file approach.
While it is important to analyse the patterns it is not right to expose the units which are not doing justice. We should come back to specific units with more specific comments as the intention is not to find faults and demonize.If the forecasting process has to take off it needs a lot of focus and detail working. We cannot be accurate to the last digit but we cannot be wrong from the beginning for want of efforts.
So if companies are doing forecasting as rituals it is time to take a review as we might be swinging from an overforecasting mode in early last year, to underforecasting mode this year. More importantly average forecasting mode is a sort of middle path which will be inaccurate in most cases. Rituals can become strong processes with strong commitment and focus. It is for leaders to make it happen.

No comments:

About Me

My photo
Gothenburg, Sweden
Still finding introspecting to find who am I? Waiting for a Guru!