Side effects of deliverability issues

Jeff Mattes makes an interesting point in a post and article on market research companies who conduct email surveys.

This point is that such surveys need to account for the proportion of emails that fail to get delivered when evaluating sample size and bias.

If you assume half the people who get the survey mail will respond and you need 500 responses, sending 1000 emails isn't enough. Because maybe only 800 emails actually get through to the inbox...giving you 400 responses.

Plus selective blocking by ISPs and webmail services means particular segments of your recipient list (say all Gmail users) get excluded from the survey, introducing bias.

But what has this got to do with your average email marketer?

Jeff's warnings and advice apply equally when you are conducting test campaigns.

If you send out a test email to a small sample of your address list, don't forget that a proportion of that test list will never see the email. So the real number of test email addresses is less than you think.

More on deliverability and statistics | Tags: email marketing testing, email marketing statistics, email deliverability

Permalink | June 08, 2007 | 0 comment(s)
Get posts like this: as an RSS feed | biweekly email | via Twitter

0 Comments:

Comments closed for this post