Invitation on arrival: the rationale

Why do we intercept visitors only on arrival to a website? Why is this method the only sampling option available for the 4Q online survey? Why not intercept visitors once they've clicked deeper into the site architecture--in other words, why not intercept visitors who are more loyal, more engaged?

Sound random sampling, within the context of an online survey, requires that everyone in a population (the population in this case is the total number of visitors to your site) has an equal probability of being selected. Those that spend more time on a website or visit more pages are almost unquestionably different from those that spend less time on websites or visit fewer pages. Yes, they may be your brand champions, yes they may be highly engaged, and yes it might flatter your professional ego to hear them declaim about their love for your brand, but, ultimately, a sample biased towards them will not be representative of your site population at large--it will not include all-important feedback from your bounces, your drive-by traffic, etc. Chances are, feedback from the latter group will be less positive, but it will be chock-full of low hanging fixes you can make to migrate these people from drive-by status to engaged status.

To get at that representative cross-section of your visitors base--in all its diverse, multifaceted glory--we need probability theory to be working on our side. Directional data from your highly-engaged, high-pages viewed segments certainly has a role to play, but representative sampling observations start with giving every visitor a chance to speak his/her mind, and that means we have to randomly invite 4Q online survey respondents at the only gate that everyone passes through--the front door.