Cash Forecasting Accuracy (#187)
/In this podcast episode, we discuss how to improve the accuracy of the cash forecast. Key points made are noted below.
The Need for a Cash Forecast
The company controller or bookkeeper might think that the main work product of the accounting department is the monthly financial statements, and that’s generally true. However. In a smaller organization, the accountant is also asked to create a cash forecast, which probably comes out once a week.
The management team depends on that forecast to figure out how much they can spend, and invest, or need to fund raise. So, what’s their opinion of the accountant’s skill level if an inaccurate cash forecast comes out every single week?
Cash Forecast Accuracy Enhancements
And that’s why we need to focus on improving the accuracy of the cash forecast. There’re a couple of good techniques for gradually improving the forecast. The first one is to limit the time period over which there is a forecast. If you’re really shooting for a high level of prediction accuracy, only forecast for one month out. Within that time range, you can use the company’s existing aged accounts receivable report to estimate when cash is coming in. The same goes for the aged payables report for predicting when cash goes out. Beyond one month, you’re guessing at the amount of sales that will be generated and you’re guessing at the amount of payables. So basically, there’s an immediate accuracy drop off one month out.
If the management team really insists on a longer prediction period, then go ahead and give it to them. But do some re-education regarding what I just said, so that they understand the issue. And on top of that, put a thick vertical line down the cash forecast that separates the first four weeks of cash forecast from anything coming after it, so there’s a built-in reminder that everything to the right of that line is suspect.
Next up, improve your accuracy within that four-week forecasting period with the 80/20 rule, which is the Pareto Principle. This means individually including in the forecast the 20% of receivables and payables that make up 80% of the cash flows. These items are so big that you have to include your best estimate for each one, or else there’s a major risk that the forecast in total will be wrong. Obviously, that’s a lot of work, but there’s no way to arrive at a high-quality forecast unless you put this kind of effort into it.
As for the remaining 80% of receivables and payables, it’s usually OK to spread it evenly across the four weeks in the forecast.
The next thing to do is install a feedback loop. Keep a copy of each weekly cash forecast, and compare it to actual results as soon as each week is over. This means really digging into the individual cash receipts and cash disbursements, to see where the differences are. And start taking notes. This is where you can really boost the forecast accuracy over time. Chances are, a lot of issues are going to appear.
For example, certain customers have a pattern of paying a little later or earlier than what you expected, or maybe there’s a more general pattern, like all customers paying later when a national holiday interferes with their payment cycles.
Or, you find that certain types of payments were completely missed in the forecast. For example, there may be a property tax payment, which only happens once a year, or a quarterly dividend payment, or annual pay raises went through for an entire department. Or, certain payroll taxes are higher in the beginning of the calendar year, because they’re capped. A couple of actions need to come out of this. One is to start developing an annual calendar of payment events, and mark on it when you expect each event to take place. Review this calendar every time you update the cash forecast.
And another outcome is to start developing a network within the company that feeds you information for the forecast. For example, the corporate secretary knows when the board of directors has approved a dividend. Or, the human resources director knows when pay raises are scheduled to begin, and for how much.
Over time, you’ll gradually figure out the nuances of the cash flows. But even then, there’re always changes going on, so the investigation never stops. Customers change their payment patterns, the company switches to different suppliers who have different payment terms, maybe there’s an acquisition, or fixed assets are purchased. Who knows.
The main point is that you have to maintain an inquiring attitude about the cash forecast, because actual results are always going to differ in some respects from what you thought would happen.
Now, an additional issue is to put into the forecast management’s reaction to the forecast. For example, if there’s a clear cash shortfall projected, then there may be a requirement to hold off on paying certain supplier invoices for a week, or maybe there’ll be a commitment from an investor to give the company a short-term loan. These adjustments need to go into the forecast, which means that issuing the forecast is really a two-stage process. Version one goes out, and then adjustments to it become version two. You should retain version two for comparison purposes when you do the feedback loop.
And a final issue is that, if management issues a mandate to delay payments, that you follow through and make sure that this happens. Otherwise, your own inaction will cause a forecast variance.
So what we need for better cash forecasting is to expand the level of forecasting detail, but only for a four-week period, create a feedback loop, install a calendar of cash-related activities, and set up a network that feeds you information about cash-related events. And make sure that you follow through on any reactions to the forecast.