We can thank modern business management guru Peter Drucker for coming up with the idea of KPIs and the fundamental maxim, “If you can’t measure it, you can’t improve it.”
That observation is as true for cash forecasting as it is for anything else. If you want to improve your cash forecasting, you need to determine what forecast accuracy KPI formulas you’re going to use as your yardstick. There are a lot of different KPIs you could use, ranging from simple calculations to quite complex ones. But the most common and useful KPIs for tracking and improving forecast accuracy are the following four:
Before getting into how you calculate these four forecast accuracy KPIs, let’s review what cash forecasting is, why you need to undertake cash forecasting and the challenges of this critical task.
Cash forecasting, or cash flow forecasting, is a planning tool. Instead of presenting a summary of current cash positions, cash forecasting looks ahead and estimates a company’s future inflows and outflows of cash across all segments over a specific period of time. No matter how well or poorly a company is performing, this report is essential for making decisions about how to manage working capital, funding and investing.
Creating a cash forecast with a meaningful level of accuracy is challenging, but it’s necessary so a business can avoid the risk of being short on cash while at the same time not having too much in reserves. Using forecast accuracy KPI formulas will make the task less challenging and give you hard data to compare your forecasts over time.
Cash forecasts that look at the next 30 to 90 days are considered short-term and are used to identify excess cash or funding needs in the immediate future. The next time frame is a medium-term forecast and covers an interval between one month and one year ahead. And a long-term cash flow forecast looks at expected revenue and expenses further into the future, sometimes as far out as five years. Keep in mind that the further out the time horizon of a cash forecast is, the less accurate it will be.
Much like a captain of a ship uses navigational tools to steer through troubled waters, business leaders need to use cash forecasts to make better informed decisions for optimal cash management and liquidity planning. A forecast that predicts a negative cash flow can be taken as a warning sign and prompt a change of course. Forecasts also help companies prepare for specific situations such as a month that has an extra payday. For startups that need to be careful about running out of cash (as that’s the number one reason they fail), cash forecasts can help them monitor their burn rate.
That’s why it’s crucial to ensure that your cash forecast formulas are as accurate as they can possibly be. When you start measuring the accuracy of your cash forecasts–using the forecast accuracy KPIs we’ll explain below–you’ll be able to improve them as time goes on.
Whether you use the direct or indirect method of building a cash flow forecast, here are some additional ways that forecasts can guide your management decisions:
Even though the positives of cash forecasting far outweigh the negatives, you should be aware of the drawbacks of going through the process of building a cash flow forecast. It is time-consuming, especially if your finance team has to manually gather and prepare the data for presentation. CFOs and executive teams want access to accurate and reliable information with flexibility in the model to test any assumptions. Then there’s the fact that frequently a forecast is inaccurate–again this is more common when the forecast formula is completed manually. You can end up with errors and inaccuracies if you:
Considering how many critical business decisions potentially will be made based on your cash forecasts, it’s imperative that you track their accuracy. Of course, forecasts will never perfectly match reality because inevitably some data will be missing and the market will present some surprises. But using the following KPIs should help you monitor the accuracy of your cash forecasts and measure forecast accuracy.
We’ll begin with the most straightforward KPI, the forecast error. It shows you the difference between your actual cash positions (Dt) and your cash projections (Ft) for the selected period. You can calculate this KPI with the following formula:
Forecast Error: 1 – [ABS (Dt – Ft) / Dt]
Dt = actual cash for period t
Ft = forecasted cash for period t
Your answer will be a percentage, with 100% representing a perfect forecast. Obviously, a forecast accuracy of 100% won’t show up very often, but anything above 70% is considered acceptable. The Forecast Error KPI is useful, but it only gives you information about the absolute value of your error, so make sure to use other forecast accuracy KPIs along with it.
This simple KPI gives insight about whether or not your forecasts tend to continue in one direction. By using this KPI, you’ll learn that you are either over- or under- forecasting most of the time.
Bias: [∑ (Dt – Ft)] / n
Dt = actual cash for period t
Ft = forecast cash for period t
n = number of forecast errors
A bias of zero indicates that there is no bias. For a period of 24 observations, if the bias is over 4 you’re most likely under-forecasting. A bias less than -4 would show you tend to over-forecast.
This is the KPI you’ll probably use the most because it’s quite simple and easy to understand. As a percentage of relative error, the forecast errors it shows are a percentage of actual observations.
MAPE: ∑ |Et / Dt |/n * 100
Dt = actual cash for period t
Et = forecast error for period t
N = number of forecast errors
You have to be careful using MAPE because it doesn’t work well in every situation. For example, if your actual cash number is low, MAPE can be misleading. Another situation is if you ever have a time period when the actual cash is zero–a number not easily divided. To avoid the first problem of having low actual cash numbers, you can use WMAPE.
If you know that it’s more important to predict some time periods with a higher level of accuracy, then that’s the time to use WMAPE. You can give those periods more weight when you’re calculating the error.
WMAPE: ∑(|Dt-Ft|) / ∑(Dt)
Dt = actual observation for period t
Ft = forecast for period t
Keep in mind that this KPI, although helpful, is too challenging to use if you have several time periods to create forecasts for and you have to weight the time periods manually. Automation will be needed.
Predicting cash flow at any given point in time is difficult without using a data-driven process. And without measuring and comparing your forecast formulas, it’s impossible to improve their accuracy. Since having accurate forecasts is essential to making informed decisions about current or future spend, growth and investing, it’s clear that ensuring your cash forecasts continue to improve should be a priority for any business.
Producing an accurate forecast is easier with an integrated all-in-one cash management tool like Payference. See how we can help by scheduling a demo today!