At this point it is useful to explore the differences between the last two runs that were made. With the two runs loaded, select Customers into the Workbench and click on the Causes Strip graph:
Now we double click on completions and again click on the Causes Strip graph. A useful path to follow is: Customers, completions, new shipments, new orders, committals, customer with non customer contacts. This gives you the final graph:
Customer prevalence stays much lower for much longer. Because prevalence is low, there is much more limited contact, and therefore fewer people are hearing about and deciding to purchase the product.
We want to focus in on the first year to understand what triggers the lower customer prevalence. Restrict the time axis to display behavior from about year 0 to year 1 (use the Time Axis tab of the Control Panel to change the time range then create a new graph).
Click on the Causes Strip graph again. Output for the narrower time range is displayed:
With this narrower time range still selected, trace the causes of customer prevalence. A useful path is customer prevalence, Customers, completions, new shipments, total shipments. When we get to the final Causes Strip graph we have:
This is the new connection that did not exist in the previous model — total shipments now depends on production, and production takes time to get up to speed due to the delays in perceiving the growth in orders and acquiring capacity.
The decreased production means that fewer people have the product, and this in turn means that demand takes longer to build. In a sense this is good, since it decreases the excess capacity overshoot as the market saturates. From a competitive standpoint it is a problem, and this is explored in Chapter 6.
