Weekly Forecasting Tips
Every week, we send out our Forecasting Tip of the week. It is a short email, with the tip of the week and an explanation. Sometimes the tip is brilliant, sometimes it is just good old common sense, and sometimes the tip is so simple that you may already be doing it. The email takes less than 2 minutes to read and once you sign up you will start receiving the tip every week. To stop receiving the email, simply hit the unsubscribe button and we will remove your name immediately.
For Email Marketing you can trust
Here are a couple of examples:
TIP 23 – “self correcting” assumptions.
"Self correcting" is an interesting idea. What if your forecasting model was self correcting? Imagine if your model fine-tuned your figures each month, learning from the actual data that you linked it to, so that your forecasts became more accurate. Does that sound like a dream or a nightmare?
If your model has a multiplier assumption within it somewhere, and it really doesn't matter which industry you are in for this to work, and that figure is used to create an important part of your forecast, then the biggest challenge is getting that figure right in your forecast. It could be wastage in manufacturing, or staff turn-over or retention or churn of customers in a services company, or drop-out rates at a University. It really doesn't matter. Now imagine that each month, the system worked out what the assumption figure should have been to have created a more accurate forecast. It then changes the assumption figure to a more realistic figure in your forecast.
How that figure is changed would be up to you. You could average out the last X months' figures, or if there is a linear change, you could extrapolate forward using the same linear trend, or you could take the average trend but apply it to a seasonally adjusted profile.
By now, you will either be very excited or very scared. Personally, I do a mixture of all of the above, and then I review the suggestions, plus the results they produce in my forecast, before manually adjusting the figure to be what I think it should be. I love technology, but I trust humans to spot the flaw, or to know when things getting better or worse start to get better or worse in a slower way.
Tip of the Week: Use technology - and then check your figures.
If your model has a multiplier assumption within it somewhere, and it really doesn't matter which industry you are in for this to work, and that figure is used to create an important part of your forecast, then the biggest challenge is getting that figure right in your forecast. It could be wastage in manufacturing, or staff turn-over or retention or churn of customers in a services company, or drop-out rates at a University. It really doesn't matter. Now imagine that each month, the system worked out what the assumption figure should have been to have created a more accurate forecast. It then changes the assumption figure to a more realistic figure in your forecast.
How that figure is changed would be up to you. You could average out the last X months' figures, or if there is a linear change, you could extrapolate forward using the same linear trend, or you could take the average trend but apply it to a seasonally adjusted profile.
By now, you will either be very excited or very scared. Personally, I do a mixture of all of the above, and then I review the suggestions, plus the results they produce in my forecast, before manually adjusting the figure to be what I think it should be. I love technology, but I trust humans to spot the flaw, or to know when things getting better or worse start to get better or worse in a slower way.
Tip of the Week: Use technology - and then check your figures.
TIP 36 – measure twice, cut once.
I was doing a spot of DIY on Saturday, eagerly watched by my 6 year old son. He hasn't had much opportunity to observe me using my toolkit since I'm not very accomplished at DIY so I either avoid doing any or avoid doing any with an audience. I still have a sneaking suspicion that spirit-levels are a form of witchcraft.
Still, I suddenly found myself trotting out clever advice that I guess I had heard my grandfather tell me. The most stunning advice, from a woodworking point of view, was "Measure twice, cut once". Now over the years I confess to having ignored this little gem as time and time again I have measured once and had to cut more than twice as I found that I had measured incorrectly. Since I am confessing, there were quite a few journeys back to the DIY shop to buy additional materials to replace the ones I has mis-cut in the first place. With all the money I have wasted, I could have ..... (bought more beer springs to mind but doesn't sound very inspiring to anyone other than me).
In forecasting, the measure twice, cut once rule should also be your guide. Don't be too hasty to make big decisions based on one month's figures, or one quarters forecast. That doesn't necessarily mean wait for another month or quarter before taking action, but it could mean taking a few more measurements, testing out your plan on a few knowledgeable colleagues and doing a bit more what-if analysis before launching a new strategy or making big changes.
Finally, use this phrase to double or triple test before making headcount cuts. Just like a bit of wood, it is easy to cut, but next to impossible to stick bits back on if the cut was wrong. The hurt feelings, the damaged pride and general lowering of moral will all take their toll, and I'm guessing the same will be true with staff redundancies too.
Tip of the week: Measure twice, cut once.
Remember - if you sign up for our weekly tips, you can unsubscribe at any time if they no longer interest you.
For Email Marketing you can trust
