This post continues on with the question posed two posts earlier, “Does the Alexa ranking influence the amount of comments that a blog will receive?” or “Does the ranking help make a successful blog?”
As you may or may not know, Alexa rankings are like golf scores. A lower number is better, indicating rank among approximately the top 10 million blogs or websites.
Let’s look at the data.
The Results of My Tedious Work
Having done many of these analyses at work, and as can be plainly seen in the table from Part 2 (which is reposted here for your convenience), I can tell you that there appears to be little correlation between traffic rank and either the total number of comments or even the average number of comments per post.
|
Alexa Rank |
Number of Comments |
Number of Posts |
Average Number of Comments Per Post |
|
| web-strategist.com |
42,610 |
276 |
20 |
13.80 |
| psfk.com |
72,817 |
123 |
99 |
1.24 |
| doshdosh.com |
10,181 |
258 |
4 |
64.50 |
| mashable.com |
2,048 |
1900 |
315 |
6.03 |
| conversationagent.com |
166,142 |
91 |
13 |
7.00 |
| coca-colaconversations.com |
949,014 |
12 |
9 |
1.33 |
| Jnjbtw.com |
1,866,554 |
4 |
3 |
1.33 |
| stonyfield.typepad.com/babybabble/ |
2,594,348 |
6 |
5 |
1.20 |
| blog.delta.com/ |
2,899 |
21 |
7 |
3.00 |
| fastlane.gmblogs.com/ |
224,713 |
34 |
3 |
11.33 |
| 1000words.kodak.com/ |
6,194 |
70 |
11 |
6.36 |
| econtalk.org |
847,199 |
42 |
2 |
21.00 |
| blog.acton.org/ |
353,567 |
51 |
27 |
1.89 |
| wishfulthinking.co.uk/blog/ |
563,848 |
14 |
4 |
3.50 |
| punny.org |
191,578 |
59 |
6 |
9.83 |
| brandstrategy.wordpress.com |
2,587,030 |
0 |
3 |
0.00 |
| wallstreetexaminer.com |
281,341 |
35 |
47 |
0.74 |
| getrichslowly.org/blog |
25,545 |
1299 |
24 |
54.13 |
| natewhitehill.com |
128,358 |
8 |
2 |
4.00 |
| careerramblings.com |
938,310 |
0 |
1 |
0.00 |
| lifehack.org |
15,369 |
263 |
24 |
10.96 |
| gapingvoid.com |
46,241 |
21 |
2 |
10.50 |
For example, econtalk.org has an Alexa of 847,199 with 42 comments and averages 21 comments per post. This rank is not particularly high as things go in the blogosphere. So, let’s pick a blog with a much higher Alexa on which I could expect more comments. Psfk.com has an Alexa of 72,817 with 123 comments BUT, Psfk.com has 99 posts for the period studied whereas econtalk.org had only 2 posts. More posts, more comments. That’s a reasonable rule. So, here the answer to the question doesn’t seem to be yes.
I could go on like this, picking out various combinations. Certainly you may pick your own combinations to study. But what really matters is how the whole sample relates to itself. In order for the answer to my question to be yes, there must be an internal consistency within the sample. Finding that internal consistency is where that handy little tool, the coefficient of correlation, comes in.
I ran correlation coefficients for the following combinations of number sets.
1. Alexa Rank to Number of Comments
2. Alexa Rank to Average Number of Comments per Post
And what did I find? Well, let’s say I think you can say the answer to my question is no.
| Measurement |
Correlation Coefficient |
| Alexa Rank to Number of Comments |
-0.28 |
| Alexa Rank to Average Comments per Post |
-0.31 |
For those readers who know something about correlation coefficients, you can see that the above pretty much means that there is no relationship between the sets of numbers measured.
For those readers with no knowledge of correlation coefficients, please let me give a short explanation. A correlation coefficient can range from -1.0 to +1.0. This statistic measures the change relationship between two number sets, or factors. The sign indicates the direction of the change, while the number indicates the strength of the change.
So, a positive sign means that an increase in one factor will cause an increase in the other factor. While a negative sign means that an increase in one factor will cause a decrease in the other factor. The number represents a percentage, think of the 1 as 100%, showing how much change in the relationship there is.
Generally speaking, and there are of course exceptions, unless a correlation coefficient is over +0.70 or under -0.70, it is regarded as statistically meaningless, or that there is no relationship between the two factors.
So our coefficients show that, for our sample, there is no relationship between the traffic level and the number of comments or the average number of comments per post.
What This Means
Since there is no relationship between the Alexa ranking and the Number of Comments or the Average Comments per Post, I can say to those analysts studying their competitor’s blogs that Alexa is not going to help you assess your competitor’s level of blog success, as I defined it in Part 1 of this post series.
Conclusion
So for you, as a competitive intelligence manager, if you’re using Alexa even as a cursory measure of whether a blog is successful for your competitor, you might like to try looking at a different measure. I know I am.
