Empirical Evidence for the Role of the Domain Name Itself in Website Performance by Karan Girotra & Karl T. Ulrich [INSEAD Working Paper]

Abstract: This paper provides the first large-scale empirical evidence of the association between specific properties of internet domain names and website performance. We analyze over one million internet domain names, linking their phonological and morphological attributes to the realized demand for their associated websites. We test hypotheses related to how the names sound, how they look, their ease of recall, and the likelihood that they will be typed correctly.
We find that certain attributes of names are associated meaningfully and significantly with the demand realized by a website. The websites with the highest demand have names that are short, include dictionary words, avoid punctuation symbols, and use numerals. The use of phonemes associated with disgust is negatively associated with performance for most websites, but positively associated with performance for adult sites. Some of these results from the on-line world are likely to hold off line, while some are not. These findings can be used in conjunction with other criteria as part of the selection process for names.
One example of the findings in the paper is:
“We find that shorter domains (measured as number of characters or syllables) perform better. This suggests that, in contrast to the literature that found recall benefits of longer names in the offline world, with respect to domain names, the increased navigational yield effect of shorter names may exceed any benefits to recall of longer names. A one standard deviation increase in the length of a domain name measured as number of characters (number of syllables) is associated with a decline in site rank of 7.34% (2.99%). Alternately, an extra character or syllable is associated with a 1-2% lower rank.”
To download this INSEAD Working Paper by Karan Girotra & Karl T. Ulrich in full, see:
ssrn.com/abstract=1714992