“I don’t need the exact figure. Just give me the ballpark number.”

This is how I sometimes do business when I am trying to buy a new car. When I am early on in deciding which car to buy, knowing that one of the candidates is about $25,000 and the other one is about $40,000 is enough information for me. The ballpark number is a useful approximation for my initial purpose. (Later I will bargain about the exact car and sales price.)

In competitive intelligence, we are often asked to assign a number to something a competitor is doing.

For instance, our management might want to know how much research and development money has been spent on the latest product from our competitor. This isn’t a number that most companies will report publicly. So what do we do? Give up? No, rather we fall back on the article of competitive intelligence faith that there is always an ethical way to give a good answer.

We find a credible way to estimate the number.

We might start with the company wide R&D number (which typically is reported). Then, we apportion it among the product lines by tracking how many new product introductions have been made in the year. A further refinement is made by overlaying the typically product cycles (e.g., 24 months from conception to marketing). That suggests that the product introduced this year was partially funded by R&D funds from previous years. Next, knowing something about where the new product fits in the product line helps us understand if it is a wholly new product or a derivative from a previous product. (Derivative products will usually require less R&D dollars than completely new products.) It is often straightforward to estimate the percentage of new design or features in any product. Putting all of this together means that we can report a ballpark number to management about the R&D money spent on a new product.

Sometimes the method is as useful as the answer.

Working through the previous example illustrates how the method teases out information and assumptions. This clear thinking has two main benefits. First, it answers the inevitable question from management about how you got to the answer. Second, and more importantly, it lays bare the reasoning in a way that allows others to challenge and improve your assumptions. When a CI professional manages this well, the ownership of the answer passes from the CI person to management. This is how it should be once they (management) are equipped with the right information.

The method also conveys the risk in the estimate.

A recent article in BusinessWeek entitled Digging Up Amazon’s Numbers is a wonderful example of a useful approximation. According to the article, Amazon is generally unhelpful in providing information to analysts that track the company about future prospects. So, Marianne Wolk (analyst at Susquehanna Financial Group), used warehouse information reported by Amazon to create a leading indicator of their future sales. That is, if Amazon increased their warehouse square footage, then that meant they were expecting higher sales. Correspondingly, a decrease in warehouse space meant declining sales expectations. This approximation will have to be tested over time to validate its usefulness. Nevertheless, it is a clever way to link something that is known with something that one wishes to know.

Public companies are sieves of information.

One recurring realization is that public companies cannot help but signal their intentions. The challenge is to use what is public to estimate what is not public. We don’t have to arrive at exact numbers. Most of the time a useful approximation (with the methods and assumptions described) will be quite valuable to management.

Do you agree?