February
17
2007
Well, it’s not particularly scientific, but the results are in. The poll was quite simple: Ask my readers how many of Eric’s books they currently own – zero, one, two, or all three. I ran the poll from Feb 11 to Feb 16, 2007 (well 17 if you count this morning) and had 23 respondents. Here’s the breakdown of responses:
- 39.13% of respondents own all 3 of Eric’s books
- 39.13% of respondents own two of Eric’s books
Channeling a marketing director I used to work with, I could say, my site has a top two box score of 78.3%. That’s pretty darn good although not the nirvana of 80+%.
- 13.04% of respondents own just one of Eric’s books
- 8.7% of respondents did not own a single book (maybe they do now
)
- Respondents are 4.5 times more likely to own two or three books than they are to own none
- Respondents are 3 times more likely to own two or more books than they are to own just one
So when Eric complains that none of my referrals are purchasing books, I can say, “Fine, write a new one, most of my readers already own the others already.”
If we combine respondents that own one, two, or three of the books we’re talking about 91.3% of all respondents. In other words, you dear readers are not the target for Eric’s book selling any more unless he writes a new one.
A big thanks to everyone who took the poll!
Clint, excellent poll and as I suspected, your learned audience is already full of (hopefully satisfied) book owners. The only question I would ask is how large of a sample did you get relative to your total readership?
I’m not at my Workstation right now but I think you have more than 33 readers. I guess I have to wonder if the non-responding readers were uninterested in the poll, didn’t respond to the title of the post, or are just sick of reading about me in the blogosphere of late (I am a bit over-exposed due to the AMA event …)
Anywhoway, have a nice long weekend!
Sample size, depends on your point of view…Feedburner says I have 291 readers so the sample size there is just under 8%. However, if I look at visitors to the site, HBX says I have 140 weekly visitors so far so the sample rate is 16.4%. On a pure percentage basis, the sample size is probably ok, but since we both know the sample doesn’t really have any statistical rigor (random sample, etc) take it for what it’s worth.
I was feeling pretty lazy when I put up the poll so I didn’t hack the MajikWidget code in order to segment around respondents vs. non respondents. Therefore, I can’t prove the second part of my hypothesis, which, as we discussed, is whether or not the readers who were predisposed to not answer were those who find their way here through search engine referrals – just over half my visits this week came from search keyword referrals and a quick eyeball of what keywords were used tells me that the majority of my search traffic comes from excel and/or report design based questions and not web analytics questions.
I can tell that click-through from the feed to the site was relatively low for the poll (27 for the last 7 days) Does that agree with what you see in Workstation?
[...] exhibiting highly-engaged behavior on my site (see late August and early September.) Given that Clint proved that conversion is a poor measure of success when trying to evaluate traffic from other… I think visitor engagement is useful for examining the non-revenue value of referring sources [...]