How to forecast sales more accurately
Hi there,
This post is all about forecasting sales more accurately. It's not too hard to get your head around and we'll break the content down into 8 steps that'll walk you through it. Let's start with why you should forecast your sales more accurately so you can plan better for success.
How to forecast sales more accurately.
Sales forecasting is important because it allows you to plan your sales & marketing activity and gives you a good idea of how the business will perform over time. If you get your forecast right, you can invest in the right marketing channels, and recruit the right sales people at the right time of year. Getting it wrong can be costly as you may end up hiring staff, or wasting time on unprofitable marketing campaigns. Successful sales forecasting isn't that hard, especially when compared to other business tasks like managing finances, managing people etc.
The accuracy of forecasts are usually measured by their "error". How far off they are from actual results. Most organisations start by forecasting sales at the end of each quarter. Which is fine for a start. However, you don't want a sales team that are continually learning from one bad forecast after another, so the key is to forecast on an accurate cadence. You don't need to use monthly forecasts if you're not in a business where other parts of your business are automated, but it's still important that you're accurate.
ACTUAL SALES VS. FORECAST SALES
In this example we'll be using weekly forecasts and we'll compare actual results against the actual daily numbers. So, each day we'll compare the forecast to what actually happens.
Looking at these charts let's consider a couple of questions:
How can we better forecast our sales? How can we improve accuracy? What should we do if actual results are better than forecasts? What should we do if actual results are worse than forecasts?
We'll now look at each of these questions one by one.
Note : All charts have been created in Google Sheets and then exported as images for this post. You can download a copy from here .
1) How can we better forecast our sales?
In order to better forecast your sales you need to understand where your business is weak or strong relative to the competition. For example, if you have a high sales volume, forecast higher. If you have a low sales volume forecast lower. If your sales are predictable and relatively stable you can use forecasts for planning purposes. But if your business changes rapidly or is very unpredictable then it's better to not forecast.
2) How can we improve accuracy?
In this example we'll be forecasting sales for the week ending the 28 day of February 2017 . We'll take the average of all the weekly forecasts from the month (28 days) and compare those averages against actual results.
Note : All charts have been created in Google Sheets and then exported as images for this post. You can download a copy from here .
Let's zoom in on where actual results and forecasts diverge.
Zooming in we can see that the biggest difference is on the last day of the month . It's clear from this view that the forecast is a couple of units short. Let's now look at how accurate forecasts are for each week of the month.
How can we use this chart? This chart shows us a view of how accurate our forecasts were for each week, but it doesn't show us why or what to do about it. Let's take a look at forecasting accuracy over time using an error vs. time graph.
How can we use this graph? An error vs. time graph shows us the amount of error between the forecast and actual results for each period of time. This is a valuable tool for forecasting accuracy because it helps us to identify specific trends or patterns. In this case, we can see that our forecasts are consistently overshooting actual sales figures. So, how can we improve accuracy? We could increase our unit forecast or decrease it depending on what's happening around us.
Over the week ending the 28th of February 2017 we forecast an average of 5,611 units per week which is close to what happened. However if you compare these results against other weeks you'll notice that we're consistently overshooting sales volumes especially on the last day of the month. Our immediate reaction might be to raise our unit forecast for the week ending the 28th of February. But should we?
We'd need to understand why sales are overshooting and how much overshooting is acceptable. For example if we get a strong delivery from our supplier on the Friday it could overshoot by a few hundred units and not cause us problems. On the other hand getting a poor delivery from our supplier could leave us with sales that are lower than planned, or worse in some cases, way lower than expected. So, let's look at what happened in the previous month .
This data shows us that the demand in February was low. March will be a different story but we'll need to take into account normal seasonality and not just what happened this month .
3) What should we do if actual results are better than forecasts?
Most people start by forecasting sales at the end of each quarter because it's a good "warm up" exercise for forecasting. If you're expecting great results from Q1, the after-effects of Christmas result in a natural boost to demand. This can be hugely beneficial, even if sales for Q4 are usually very good. So, forecasting on an accurate cadence means using accurate forecasts for Q1 and keeping an eye on them throughout the year.
4) What should we do if actual results are worse than forecasts?
If actual sales are worse than the forecast our immediate reaction is to slash forecasts. This can be dangerous so we need to understand why actual results were worse than forecast . If a product launch went badly or a competitor launched a new product which was much better, you may have overestimated the demand for your product. Or, lack of demand for your product could prevent you from making sales. So, it's important that you understand both short-term and longer-term causes of lower than expected sales.
5) How can we use the charts to improve accuracy?
One of the biggest mistakes that can be made by business owners is that they don't understand why actual results are worse than forecast. This is an area where data and pattern analysis can help greatly. For example, we might find that demand for our products in January was unusually low which we didn't take into account because it happened last year. Another mistake is to apply a "make or break" approach to forecasting only if sales are below expectation. Sometimes it's best to wait for a few days before slashing forecasts or raising them too high if you're fixated on getting everything right on the first try.
Conclusion
Understanding how accurate your forecasts are and how to improve them will help you to create better planning decisions. You need to be able to compare your forecast with actual results and identify any discrepancies. This is where the charts we used in this post help, but there's a lot more that you could do with them. For example, you can get a view of your sales by date range or by category of product or event which will give you additional insights into how your business is performing.
What would you try if? Let us know in the comments below.
Read my previous article on forecasting: http://www.smallbusinessbriefs.
Post a Comment