What is the impact of patient –or physician oriented marketing on the sales of prescription drugs ?

March 29th, 2010

Since the rules for patient-oriented marketing of prescription drugs have been relaxed in the US in 1997, direct-to-consumer (DTC) expenditures have more than tripled. Nevertheless physician oriented marketing (personal selling by sales reps (the so-called detailing), advertisements in journals, other marketing activities such as seminars,…) still accounts for the largest part in the pharmaceutical marketing budgets. Both pharmaceutical companies and health care policy makers are interested in the impact of this pharmaceutical marketing on the growth this can have on the total market size of a particular category of drug. Will DTC expenditures increase the market size? What is the best marketing strategy for a brand manager?

Many studies have therefore focused on the effect of pharmaceutical marketing on the size of the market for a particular category of drugs. In most cases the growth of the market size for a particular category of drugs is compared to the total marketing expenditure in that category over all brands (the so-called category sales models). On average these types of analysis find that patient oriented marketing has a greater impact on the market size than the physician oriented marketing which is found to be small.

In [1] a different approach is taken. Instead of aggregating expenditures and sales over the different brands, a brand level model is developed taking into account the competitive dynamics between brands. With a simple example it is shown that an increase of 50% in total marketing expenditure in a category with only two brands can lead to different impacts on the market size depending on the competitive dynamics between the two brands. Some brands will be able to increase their sales figures without draining sales from their competitor (thereby increasing the market size) while others simply drain sales from their competitor without increasing the total market size. On average this can lead to a situation in which total marketing expenditure has no effect on the total (over all brands) sales figures. In [1] it is shown that the previously found low responsiveness of sales on marketing expenditure at the category level is in sharp contrast to the estimated sales effects for individual brands. Using the brand level model it is found that sales show the highest responsiveness for detailing, followed by professional journal advertising and DTC advertising. The reason that the impact of DTC seems to be less at the category level is that the substitutive effect of detailing has dominated in the past (ie draining sales from the competitor). The study also shows that the sharp rise in DTC expenditures did not lead to a jump in drug expenditures due to the competitive dynamics.

[1] M. Fischer, S. Albers, “Patient- or Physician-Oriented Marketing: what drives primary demand for prescription drugs ?”, Journal of Marketing Research, Vol XLVII, pp103-121, February 2010.

Can we predict customer lifetime value ?

March 15th, 2010

Quite often a small percentage of customers accounts for a large part of the revenues and/or profits. From this observation a company strategy could be to detect customers with a high customer life time value, i.e. customers that will generate many profits to the company in the future and give them special (marketing) attention thereby assuring not to loose this valuable “asset”. A key element in such a strategy will obviously be the prediction of customer life time value based on past observations (money spent, use of services,….). The question is of course: can we predict customer life time value?
An interesting paper on this subject was published in the Journal of Interactive Marketing [1].

In this article mainly discretionary marketing investments such as giving some customers priority based on past behavior (room upgrade, waiving of late payment fees,…) are considered as opposed to quid-pro-quo investments (e.g. loyalty programs: if you buy more, you will get more). The reason of this focus is that quid-pro-quo investments depend on actual future behavior whereas discretionary marketing investments depend on predicted future behavior (where the prediction is based on the past behavior).

It is obvious that the better you can predict customer life time value, the less “dangerous” it will be to apply discretionary marketing strategies. Clearly prediction can go wrong and someone that is predicted to be a very good customer might turn out to be not so profitable (false positive) and the other way around: someone predicted to be a normal customer might turn out to be a very good customer (false negative). It is hard to say in advance for a particular company if it is possible to predict customer lifetime values accurately or not. However, in [1] several cases are analyzed using several prediction models and surprisingly enough it turned out that for all these cases, of the actual top 20% customers 55% will be misclassified by the models (i.e. false negatives) and of the actual bottom 80% customers 15% will be misclassified (false positives). This leads to what the authors name the “20-55” rule and the “80-15” rule.

As the authors conclude, the question whether investments in discretionary marketing make sense can only be answered with a profit model using the probabilities and costs of misclassifying customers, the additional revenue generated by the investments and the cost of the investment itself. These probabilities should of course be estimated for a particular company or sector based on historical data.

[1] Edward C. Malthouse, Robert C. Blattberg, “Can we predict customer lifetime value”,  Journal of Interactive Marketing, Volume 19, Issue 1, 2005, Pages 2-16