The Paradox of Invitations

Lately I’ve been intrigued by the idea of website and web app invitations. You join a website and then gain some benefit (as does the site) by inviting a bunch of your friends. A lot of these invitation widgets access your e-mail, Facebook, or LinkedIn accounts. You click “permission to access gmail,” and all of a sudden you’re presented with a list of 300 e-mail address from which to choose.

I don’t know about you, but at that point, unless the incentive for getting a friend to join is really good or if I have someone specific in mind, I just skip the invitations. It’s the paradox of choice: I’m simply overwhelmed by the sheer number of Facebook friends or e-mail addresses from which to choose, and so I choose none.

I’m wondering if there’s a better way for a website to suggest invites to a new user. Thus I propose the following experiment:

Borrow a website that gets thousands of new users every month. Randomly rotate the format of the invitation page so that each of the following invitation formats are equally distributed (thus, when User A lands on the invitation page, he sees Scenario 1, then the next user sees Scenario 2 instead [and does not know of the existence of Scenario 1] and so on):


  1. Regular invitation page where you are given a list of all of your Facebook friends and are asked to pick as many as you’d like.
  2. Invite page where you are given a list of all your Facebook friends and are asked to pick 10 of them.
  3. Invite page where you are given a list of all your Facebook friends and are asked to pick 5 of them.
  4. Invite page where you are given a list of all your Facebook friends and are asked to pick a specific group with no quantifiable guidelines. I.e., your best friends, your immediate family members, your coworkers.

The invitations would read as follows: “Hey! I just signed up for this fascinating new site. I have the one-time option of inviting exactly 10 of my friends and colleagues, and I chose you. Check it out if you have a minute.” (Where the number “10” matches Scenario 2.)

The results I’m interested in are (a) the number of invitations per new site user (not “new site users who decides to invite people”) and (b) the number of conversions (new sign ups) as a result of the invitations.

My hypothesis is that Scenario 4 will result in the most invitations per new user because the assigned context makes it easy for our minds to jump to specific faces and names without much thought. However, I would propose that it will be a much closer contest between scenarios 3 and 4 for the number of invite signups per new user. Scenario 4 is somewhat flattering and will target people who generally trust your taste, but Scenario 3 is flattering in a different way: The quantifiable number in the invitation makes people feel like they’re in a very limited, exclusive, elite group (or maybe they’ll feel like they’re the people to whom you pass on annoying website invites, I don’t know).

I wish I could run this experiment on my blog, but I think it would be more effective if tried on a much heavier trafficked website. Thus I’ve proposed the experiment to behavioral economist Dan Ariely (isn’t it cool that we live in a world where you can propose an idea to a bestselling author and he’ll actually read it? Perhaps even the next day?) [UPDATE: I heard back from Dan, and he’s going to keep an eye out for sites to try this method, but if you know anyone who runs a popular website with whom we could try this little experiment, contact me at]

What’s your hypothesis?