The learning curve is a mathematical management tool normally used by manufacturing companies to evaluate productivity gains caused by organizational learning.

Companies can obtain many benefits from using a theoretical learning curve in their production scheduling and general future prognosis of effectiveness.

Firstly, if learning curve improvements are not considered when creating productions schedules, the result might be idle production facilities, where both workers and machinery are not creating any value. Secondly, companies may decline additional work, because the company does not take the learning curve�s effect on production capacity into consideration. Thirdly, companies could also use the learning curve to negotiate with suppliers, who should also be able to learn over time, and hence become less expensive. Fourthly, companies may be able to give a future prognosis of the development of production costs, and thereby analyze the future effect of the organizational learning on profitability.

The examples above are however not exhaustive and companies might very well find several other good uses of the learning curve.

Calculating the learning curve.

In this article, two ways of calculating the learning curve is presented � The arithmetic approach and the logarithmic approach.

Arithmetic approach

The basic thesis of the learning curve is that every time the produced volume is doubled, labour per unit will decline by a constant rate, which is normally referred to as the learning rate.

The example above uses a learning rate of 80%. The hours used for producing is therefore 20% less every time the actual produced amount has been doubled. The problem above is however that we cannot see how much labour will be required for units that are not equal to the doubled values. For calculating labour consumption for units that are not equal to the doubled values, we will have to use the logarithmic approach.

Logarithmic approach

The logarithmic approach lets us determine the labour consumption for every Nth unit using this formula.

 

T_N = T₁(N^b)

 

Where:

T_N = Time for the Nth Unit

T₁ = Hours to produce the first unit

b = (Log of the learning rate/Log 2) = Slope of the learning curve

 

If the learning rate for a given operation is 80%, and the first produced unit took 1.000 hours, the hours needed for the thirteenth unit can be calculated as follows:

 

T_N = T₁(N^b)

T₁₃ = (1000 hours(13^b)

= (1000)(13^log0,8/log2)

= (1000)(13^-0,322) = 438 hours

 

The thirteenth produced unit will therefore take 438 hours to produce. At the end of the article, the characteristic learning curve slope is depicted in a graphical chart.