I think one should always remember that, as with responses to adverts, just because one person offers to work for me as an assassin, that doesn’t mean all the clientele of the Bollard are this way inclined.

The only measure that matters are the broader figures.  For example, if you send out an email you should be asking...

  • What percentage of people click through from the advert to your website (an indicator of the effectiveness in the email)?
  • What percentage of people go from the website to making a purchase (an indicator of the effectiveness of the website)?
  • How can I get those percentages higher?
  • And if the assassination gets bungled, am I likely to be implicated.