Friday, March 24, 2006

On Measuring the Ability to Innovate
It may help to keep a simple principle in mind

Bruce Nussbaum of Business Week provides an example of how Wall street is hopelessly amiss in the knowledge economy. Apparently, Apple Computer is being considered to be flagging in its innovativeness, because it's R&D budget has been declining as a percentage of revenues!

This got me thinking. Admittedly, analysts need yardsticks to evaluate companies, but these yardsticks must truly measure whatever is being sought to be measured, and not just the closest possible proxy. A "something is better than nothing" approach simply does not work!

The Horse Sense take on this would be that, measuring success in innovation by looking at the size of the R&D budget is like figuring out how successful a song (or a film or a book) will be by measuring how long the creator took to write it.

An approach such as the above fails because it assumes that, while measuring a company's ability, output is "proportional" to input. This may be true for attributes that are fairly stable and well-understood - for example, if you are a car manufacturer, spending more to buy raw steel of higher quality would probably be expected to result in cars that are stronger or safer. This is because most manufacturers can be expected to be pretty similar in terms of how they use the raw steel to make the car - in other words, different manufacturers wouldn't differ much on "how well" they use the raw steel.

However, such an approach breaks down when the attribute under consideration can be considered a "differentiator" for the company, as then by definition, "how well" the company uses the input becomes critical. Thus, throwing big money to recruit the best designers may not necessarily help a car manufacturer produce cars that are superbly designed - that would depend on "how well" the company uses the design talent it recruits!!

Further, the "how well" principle holds more strongly as the attribute under consideration becomes less well-understood, less well-structured, and less algorithmic. In other words, as you measure an ability that is less well-structured, the "output-is-proportional-to-input" argument grows increasingly weaker.

And the principle perhaps applies most strongly while measuring innovation, which is arguably the most arcane of corporate skills today!