RL Magazine
Technical Trends - Gedankenexperiment
by L. Bryant Underwood

Return to Menu

One of the things that Albert Einstein is noted for was the method he used to solve problems. He would approach questions by amplifying the principles associated with the problem in his mind by what he called in his native German, a gedankenexperiment or “thought experiment”. He would then reconcile the extremes into a new idea that would become the solution to a problem. Interestingly, there is a similar approach you can take in gauging your target position in pricing and where you need to transform your business versus market share in RL. The basis of this method is using a demand curve.

Information on using Demand Curves for market analysis is mostly centered on new products and commodities. A lot of this thinking is not quite as portable for use in RL. But with a little tuning of the methods and some thought from your subject matter experts, we can use demand curve analysis to learn a great deal about your business.

Real World Example:

Let’s consider a company that has a depot repair and refurbish operation that focuses on smart phones for one manufacturer. The key metrics for this company are;
Volume 53000/Mo
AUP $58.00 (Labor+ Mtls)
SGA $185K/Mo,
GM % 35%

The first step is to define the TAM (total available market h/t Steve Manning) expressed in gross units. For smart phones the repairs will be a subset of the gross returns. The gross returns can be estimated to be ~0.7% - ~1.2% of the installed base per month. For smart phones in the US, subscriber numbers recently crossed 100M units. That would equate to a gross return rate across the US of ~900K units per month. For a single brand of product with a 30% market share the TAM would be ~180K units total per month. We can instantly see that our example supplier above with 53K units/Mo, has ~29.44% of the TAM.

Next, we need to set the slope of the demand curve. For the RL business this is tough because there is so little information sharing or open benchmarking. However, since we are most interested in the slope and not the actual values, educated estimates coupled with some G2 from an informal survey over lunch or dinner will work very well. In this example I am setting the extremes at;

Market Share Total AUP 2.0% $74.50 85.0% $25.00

With the limits set, you then spread the data linearly in a spreadsheet across the extremes. You will end up with a dataset that looks something like this;

Return to Menu