Cash flow forecasting: do it better and save.

• Author: Ochs, Joyce R.

Cash flow forecasting--projecting cash flows in the short term (up to one year)--is an important financial management tool. If it is not done effectively and regularly, companies can lose substantial amounts of cash, as well as opportunity costs. Knowing a forecast is bad or unreliable often causes CFOs and treasurers to make overly cautious investment decisions or borrow too frequently and be ambushed by one surprise cash need after another.

If it's so important, why can't forecasting be done well? Good question. First, every situation is different: Companies of similar size can have substantially different cash flows in terms of transaction size, the frequencies or timing and the method or location where the cash flow occurs. In other words, forecasting cash flows must be tailored to the company, a process that can be time-consuming.

Another factor is the corporate climate. Experience shows that companies with large amounts of excess cash and virtually no short-term debt place less importance on cash forecasting than those with substantial short-term borrowing activities. This makes sense, since companies that borrow have a finite amount of short-term reserves--their short-term credit capacity--while those with short-term investments do not really see such limits.

Finally, many companies have trouble corralling the various internal sources of data needed for reliable forecasts on a regular basis. Although this is changing with the growth of enterprise resource planning (ERP) systems, it will still be some time before financial managers will be able to tap into other systems to retrieve forecasting data.

What can companies do to improve forecasting? Successful forecasting requires a good strategy combined with reliable sources of data. There are several key things to do:

* Develop a baseline. How satisfied are you and other users with all aspects of your current forecasting procedure? Are some sources harder to work with than others? Can you develop independent estimates to substitute for poor sources?

Try to work on each source directly to improve its performance. Identify key problem areas; for instance, are you asking the source to forecast something it cannot do? Watch your terminology--it's better to let sources forecast in terms they understand; you can translate, consolidate or recombine the data after you receive it.

* Match required accuracy and reliability. You'll need greater accuracy if you are trying to forecast daily...