Some Issues In Modeling Life-Cycle Component Of Markets

Some markets have strong component coming from products that need to be replaced because of burn-outs, breaks , wearing off or changes in properties that render them dangerous or useless. Automotive market is a good example for such market - tires wear off, the oil, spark-plugs and other components loos properties and so on. Another example for this type of market is the lighting market - the lamps burn out and need to be replaced. Modeling this type of market is eased because of the "mechanical" relations between entities. However the challenges the modeler faces are not negligible and could seriously damage the model quality. 

Before making some points on the challenges let me touch briefly on the three components of market. First one is the market coming from worn off of broken product. For example - a tire blows off and needs to be changed. The second is coming from replacement of products that are still good. Example - a retail store changes interior and new light sources are installed. The third component comes from the new installations - a new car is produced, new home is built and so on.

I would like to share following challenges I have identified:

Banning

This is probably the least challenge. Banning is legislation that forbids production and sale of a product with specific parameters or technology - the incandescent bulbs for example. Modeling the effect on the market involves estimation of volume that goes out and how it will happen over the time. Some models rely on the some reducing the market with some percentages based on estimated portion of banned product and expectations on diminishing over time - calculated or manually-set. I find more appropriate a model based on the absolute volume in banning year and subtracting the volume to go out due to wear off or burn outs. A serious issue with banning is the replacing product - there could be more than one viable option for replacing the banned product. The consumer preferences are expensive to obtain and could be misleading.

Penetration Of New Technology

The classical case with one new product replacing a traditional one is well studied and s-curves are in the text-book. However, I see two issues here. First is the pinning of the take-off and penetration points. The s-curves are very sensitive to these parameters and small errors could lead to significant deviations of the model. The expert opinion is usually the option for the estimates. The other challenge with new technologies comes with competing more than one new technologies. The logical approach is to compare the traditional product with each of the new technologies. However, consumer preferences to one new product over another are difficult to catch as new technologies are very close and users are not aware or do not care that much for their relative advantages. Another approach is to pack together the new product and compare all new with the traditional one and then apply a split in the "new" package.

Data And Parameters Availability

The success of a model comes from the deeper insight and detailed segmentation of the market.  The deeper insight requires greater detail in  the data and parameters and here comes the problem. I mentioned the price and uncertainty in obtaining the consumer preferences that relate to penetration of new technologies and replacement of banned ones. The other problem is the whole bunch of parameters related to the life-cycle of the products in different applications or segments - the deeper we try to go and increase the value of the model the more parameters we need. Some industries may have data in a great detail but overall the deeper you go the scarcer the data and more expensive it gets to obtain. Of course, it is matter of decision on intended model accuracy and level of detail but the value is in the detail usually.

These are the few of the challenges I have faced while working on some models. The approach for tackling the problems depends on the data availability, budget, level of detail and purpose of the model. I would be happy to discuss other problems a practitioner has faced as well as to share more specifics on my solutions.