We recently had an interesting discussion about product S-curves and the ways of controlling them. Is it possible to restart the growth of a product approaching the top of an S-curve? What can a manager do to inject energy into a fading product?

Here is what the classical product S-curve looks like:

The blue line is the total number of customers. The red line is the first derivative of the blue line, i.e. the rate of new customer acquisition. The maximum of this curve represents a regime change for a product, a change from accelerating growth to eventual decline.  

The classical S-curve is described by the Bass innovation diffusion model that makes a few unrealistic assumptions. In particular, it assumes that a product can be sold only once and that after a prospect becomes a customer, they never leave.

In reality, companies repeatedly sell product updates and license renewals to their existing customers. Customers do leave from time to time by switching to products from other vendors.

A system dynamics version of the Bass model can be found in Chapter 9 of John Sterman’s “Business Dynamics.” I modified the model to account for repeated sales and customer attrition, then used it to generate the diagrams shown below.

Next, I compared the diagrams with the empirical data collected from quarterly unit sales reports published by Apple and RIM.

There is a strong resemblance between the blue curves in these diagrams and the unit sales curves generated by the model.

Encouraged by this validation, we can make a couple of interesting observations.

First, customer retention matters. It is the difference between being an Apple and a Blackberry.

Second, even with the perfect retention, quarterly sales reach a peak and then decline precipitously before landing on a period of relative stability. This decline can last for months or even years. Look at the iPad picture.

This is worrisome. Very few sales or marketing leaders can survive 12 months of declining quarterly sales. Can a new leader make a difference?

To model the situation, I added a 100% step increase to marketing effectiveness at month 36 and smoothed the change over a 12-month period to account for the time it takes to implement the improvements. Here is what I got:

By doubling marketing effectiveness, we managed to reverse the decline but only temporarily. The results are somewhat counterintuitive. There must be something else that we can do.

System dynamics connoisseurs will recognize the bumps in the diagrams as manifestations of a delay in the system. Upon some reflection, this delay appears to be caused by the difference between the lengths of the product release and product sales cycles. The former is usually much longer than the latter. When a product is launched, the sales team quickly exhausts the market capacity and must wait for a new version of the product to be released before sales can resume. Hence the delay.

Paradoxically, the more successful is the sales team, and the more effective is the company’s marketing, the higher is the peak, and the steeper is the decline.

There are two ways to address the problem. One could go after a larger market to increase the time it takes to exhaust its carrying capacity. Alternatively, one could reduce the length of the product release cycle, e.g. by decentralizing decision-making and implementing an agile development process.

The success of implementing the first alternative depends on a range of exogenous factors. The second one is entirely endogenous, which makes it a better place to start for most companies.

To model the impact of the second alternative, I added a 100% step reduction in the length of the product release cycle and smoothed the effect over the next 12 months. The results are shown below.

The positive effect of a shorter product release cycle takes longer to materialize, but the cumulative impact is much more significant. On the downside, a shorter product release cycle leads to a reduction in the overall number of customers. Shorter release cycles amplify customer attrition processes, which may or may not be a problem depending on the business model.

In summary:

  1. An eventual decline in sales after a successful product launch is almost inevitable. The more successful is the launch and the more viral is the product distribution, the steeper is the decline.
  2. Improving marketing effectiveness after a sharp decline is helpful but will bring only temporary relief.
  3. Meaningful innovation at faster clock speed is essential. A decline in unit sales can be reversed by implementing a shorter product release cycle but at the cost of reducing the overall customer population.

Ping me if you’d like to get a copy of my system dynamics model.