The forecast was wrong!

A familiar phrase to all us planners, and one which we just dread hearing. Used by many in the business as an excuse for issues in performance, it is often seen by us planners as a case of ‘passing the blame’ by other areas of the business. Just think how it feels to be told that your work is always wrong? But are planners the architect of their own downfall?

Let’s step back and look at the dictionary definition of a forecast. A forecast is to ‘predict or estimate an event in the future.’ It is also ‘a calculation or estimate of future events.’ Some interesting words there – particularly ‘estimate’ and ‘predict.’ Neither gives much confidence in terms of accuracy`as they both suggest that a forecast is at best an estimation of what may happen in the future.

So why then, when asked to give a forecast of next week, next month or next year, planners are determined to provide an absolute number – coming up with figures like 1,235 or 84,256. Very price, very definite and dare I say it, numbers which do suggest that there is significant effort behind the chase for the magic answer. But let’s ask a key question – what are we forecasting? We aren’t forecasting how many detective phones will come off the production line at Samsung or Apple. We are forecasting for human behaviour – something entirely different than the results of a mechanical production line. There are so many things which can happen when we look at human behaviour and that in itself is exactly why our forecasts will be wrong.

Let’s think of some key examples of what can go wrong…

• A parent goes to make a call and their young child wakes up and calls – they are now likely to choose to call at another time

• We go to call a business, but another call arrives inbound and we totally forget about our original intention

• We’re watching TV and an advert or a news item reminds us to make the phone call

• A parcel gets damaged in transit, and is one of many in the consignment suffering a similar outcome, leads to unexpected calls

I could go on, but you see the picture. Noise exists all around us. By noise I mean that there are random events which means that deviation will happen from our ‘perfect’ plans. I will blog soon about Poisson distribution and the Maths behind that. However for now, lets see what that noise means.

Many of us have forecast accuracy targets. 5% is a popular number out there – one I have issues with as there is no scientific basis for that being the magic number. However, for now it is a good barometer of what happens in the industry today. There are 2 ways of looking at 5% – if for example we expect 100 calls then as forecasters we will be happy to get more than 95 but less than 105 calls. Some people would say that is a 10% range and for 5% we should really be between 97.5 and 102.5. However for now let’s stick with the former – we are allowed deviation by 5% either way.

It poses one question – why not tell operations the forecast is for 95-105 calls? It may not sound as impressive as simply saying that we expect exactly 100 calls, but it is definitely more accurate. Have a look at your data and see how often your actuals lands within the banding. Try mapping 3 lines of a graph – forecast + x%, forecast – x% and the actuals.

What you should is a squiggly line, often going up and down between the first two lines, occasionally going outside them. This changes the dynamics of the discussions we are having immediately and starts to show that planners are ‘in the zone.’ It’s a small change but one which I have seen have immediate results every time it is used.

How we portray the message we are trying to deliver is definitely one communication tool we should hone. I mentioned in another post that communication is a key skill for planners. That is not just verbal, it is how we deliver information in any format.

Is the confidence in our models our biggest downfall. Portray the acceptable inaccuracy levels and then focus conversations on things that matter. Get past the forecast is always wrong mindset.