Studies & Opinions
Tactical Domain Monetization Allocation: A Statistical
May 14, 2006
The paper points out the limitations of the current popular domain
monetization allocation tactic and outlines a statistical approach
to overcome them. Based on a domain name’s characteristics, it
outlines a model to predict the best monetization service provider
for the domain name.
Current Methodology & Limitations
The current popular approach is based on splitting a portfolio
of domain names among various monetization service providers,
then comparing the performance of each sub-portfolio. This approach,
however, has three limitations: it can only determine relative
performance; the comparisons are based on unlike items, like apples
and oranges; it cannot distinguish between random and systematic
performance. The first issue has been discussed elsewhere. Thus, this paper develops a statistical approach to tackle the
other two limitations.
Overcoming Current Limitations
To ensure that apples are compared to apples, domain names need
to be identified based on significant distinguishing characteristics,
such as use popularity of the implicit keywords and pay-per-click
rates. Such an approach has been used to determine premiums among domain
name marketplaces and can be easily adapted to measure return
premiums among monetization service providers.
Revenue performance over a period of time has
two components: systematic and random. Thus, historical monetization
performance can be important in predicting future performance
of the systematic component. Hence, the model is re-estimated
Using the history of estimated models, based on a domain name’s
characteristics, one can predict the service provider that is
expected to generate the highest revenue for the domain name.
Predictions based on the most recent month’s data are weighed
This analytical tool is timely, as there
are currently over ten monetization service providers and the
field is growing.