SCM Awareness: The Forecasting Process 4

Now that you hopefully understand the forecast process you can start to create one into the future to the end of the “forecast horizon”.  So, is it best to use the quantitative (numbers) method or the qualitative (personal experience & opinion) method?  My personal preference was to use a combination. I would use forecasting tools to do the hard-mathematical work and establish a base forecast to start. That could be as simple as simply copying the past month’s numbers or more complex such as linear regression etc. 

But relying blindly on forecasts that simply take historical data is akin to steering a car using only the rear-view mirror to work out where you are going next. What you are trying to achieve is insight into the future behaviour of your customers and data alone is not insight.  Behaviour is determined a number of different human factors so to determine this you need to factor in the market intelligence and other qualitative information you have gathered. In some the forecast will change little and in others it will change quite a lot.  If you understand why and can stand over the data, then that is OK. 

And now, unfortunately, some political and business realities kick in.  Forecasting is often done in two ways.  Bottom up and top down.  Bottom up is the logical way of doing it.  It is what I have shown you up to now.  You take the calculated numbers and input from the various sales teams and create a combined forecast that tells you what to plan to produce. Easy?  But that must be compared with the forecast requirements from the long-term plans of senior management. 

For example, senior management may have promised investors that the company will have $5 million sales in the next 12 months.  Money has been committed to the company on that basis.  If you produce a forecast that says sales of just $3 million then you are likely to find that senior management will take an active interest in understanding the gap and pressure will be put on the sales team to deliver an extra $2 million in sales.  Is that fair and realistic?  Well it does happen.  I have seen situations where sales are set targets and are told to meet them.  Now the easiest, and worst, thing that can happen is sales just say OK. We will up our forecast to match your numbers. But they have no intention to meet those numbers.  Now all you have done is set unrealistic targets and given the sales team 12 months to find another job before the whole thing blows up.  It is critical that when there is a mismatch between the expectations of senior management and the realities on the ground that this is communicated as early as possible so that a realistic solution can be found.

Now what I have just told you may seem totally counterproductive.  You are not likely to find this referred to in any academic book about forecasting.  But unfortunately, it is something that happens.  You could spend weeks working on a forecast only to have senior management tell you to change it to match a number they pretty much invented out of thin air 2 years ago.  It is hugely frustrating, but it is a political reality. If that happens you need to get it documented that your forecast is being overridden by senior management. If there is an inquiry in the future on what went wrong with the forecast, then you will be covered (hopefully). 

Now that you have calculated the forecast you will want to review, and sense check it vs past sales and any known information (e.g. promotions etc).  Now you may see this as being a natural part of the process of generating the forecast itself. And if you do good for you.  You are 100% correct. However, it can also be the case that people are just so tired after going through the forecast process that they do not sense check it and miss obvious errors. Mistakes can happen.  So, make sure you sense check your numbers and see that they make sense.  Once you are happy that you have not missed anything get someone else in your department to sense check the numbers.  Sometimes the forecast owner can be so close to the data that they do not see the mistakes.  It happens. It has happened to me.  So, get someone to check it.  

Once that is done you need to communicate the forecast.  There is a sequence you need to do that in. 

Firstly, share it with the sales team.  They are the people who are responsible for meeting these numbers. If there is something wrong with it or missing from it then they will spot it.  Now I am not talking about sending it out to each salesman or woman in the field.  Big organisations could have thousands of salespeople. What you should do is have a few key contacts one in each region or division or whatever the natural structure of the company is. Run the forecast by them and get their input. 

Once that is done run it by the New Products Development team (if you have one).  Make sure what they are working on is included in it. 

Then I suggest you discuss it with the finance department.  I worked in one company where the finance director and I would talk regularly about the forecast. That in a way was pre-empting the top down or bottom up issue.  If we were aligned before walking into a management meeting, there was much less chance of any big dispute about the forecast. 

Once everyone is happy with the forecast you share it with operations so they can see what they are required to produce over the next few months.  It also allows them to start to plan capacity materials etc.  Now it is likely in most manufacturing companies that the forecast owner will be part of the planning team that have to apply the forecast in real life so this may be a natural transition.

Finally, once you have a forecast and everyone is happy with it this is the time to agree and implement a regular process for reviewing the forecast performance and adjusting/extending the forecast into the future. You may already have this which is great. But if you are implementing a forecasting process for the first time in a company is it best to put the process for the future into place now while everything is fresh in your mind and the minds of others.  The first time you create the forecast it is a huge undertaking.  But if you now have done that the best thing is to put processes in place to regularly review and adjust the forecast.  It is easier to make small changes every week than to do a large forecast review once a quarter.  And once you have settled on a regular process make sure to schedule it. Put it in people’s calendars so that it happens.