The rate of technological innovation has traditionally acted like ocean waves. The rate of growth rises and falls in waves and, occasionally, several waves amplify each other causing a period of particularly strong technological growth. In the aggregate, innovation until the end of the 20th century was a simple linear model, rather like this:

In this graphic, the solid line represents the level of technology extant while the dashed line shows how a typical user will adopt the technology. One of the central assumptions is that an average user (not an early-adopter, maven, geek, or Luddite) will avoid continual adoption of new technology. Instead, they will wait for a critical user base, trendiness, or external requirement to be associated with the cutting edge technology before investing the time/effort/money into upgrading

John Seely Brown argues that this rate can no longer be accurately modeled in a linear fashion. Instead, he proposes that the rate is an exponential function, like so:

If this is true, average technology users face a developing problem similar to that which plagues inventory management specialists. Is it better to upgrade (or restock) when a determined distance between the lines exists or only to do so after a determined period of time? Imagine that the user is a firm and the technology is the operating system on their computers. Upgrading to the latest version once a year is equivalent to the periodic method. As the technology curve rises, the distance between the dotted line (representing the deployed operating system) and the latest version will increase as time goes on. This means that the firm will have to expend ever-increasing amounts of time and money on training their employees since the gap between what those employees are comfortable with and what the company wants to use widens.

On the other hand, if the firm decides to upgrade each time the distance between the lines is the same, they will find themselves having to retrain those employees with increasing frequency. Again, this is not sustainable as it entails spending a larger fraction of company time and money on training (due to frequency rather than intensity now).

However, for an individual piece of technology, the development curve is rather different. As the idea becomes more refined, the rate of technological growth slows. For example, early automobiles differed widely in all aspects of their layout, engineering, and design while the functional base of current automobiles remains fairly similar for all major brands. The development curve would look something like this:

As I said earlier, adopters of technology are rarely willing to continuously upgrade. Therefore, an ambitious adoption curve like the dashed line here is highly unlikely (and cannot be recommended as efficient.) This past weekend I had the fortune to talk to Mitchell Savage who explained how the company he works for is approaching the idea of selling a continuously evolving technology.

Essentially, if I understand the strategy correctly, one wing of the company will develop technologies in a natural, unconstrained manner. This is approximated by the solid development curve in the graph below:

The marketing/sales division will then concentrate on selling given releases of the technology. In this graph, each release is represented by the round dots. As technology is developed, stable versions are produced and sold to customers. The horizontal dotted lines between the development curve and the adoption “steps” represent time that can be spent on marketing campaigns and sales pushes while the horizontal dashed lines show the time customers can use for deployment and use. Note the concurrent dotted lines; it is possible to have a closing campaign for version 2.0 while the campaign for version 3.0 is ramping up. Note also that the versions adopted by customers (the circles) follow a more steady time/tech level growth rate.

The benefits here are many. First, the revenue flows are more evenly distributed since revenues are not directly tied to tech development . Second, customers can schedule around planned technological upgrades due the more regular length of periods. Third, marketing/sales forces have time to conceive, launch, execute, and wrap-up campaigns without worrying that a new version might be released halfway through. Finally, marketing/sales divisions have a reduced motivation to hound the developers which allows developers to concentrate on features and stability rather than appraising salespeople of the latest ideas in the pipeline.

All in all, this model seems best suited to maintaining the primacy of technology developers over the “suits-and-ties” of the enterprise. Will this become the standard for all R&D type businesses?

Tags: Marketing/Strategy