It is a study we would like to do. The plan is to select a random sample of domains for each length and query a source, e.g., Alexa, show a confidence interval by length for domain traffic.
The stats hypothesis would be to test whether length can be used as a predictor of traffic.