Website statistics (web-stats) exist to answer the question - “How many people are looking at my website?” When it comes to blogs, they suck at it.
It’s been a couple of years since I was a ’stats monkey’ - obsessively reloading a page of numbers, observing every click on my sites with such mania you would have assumed I was getting a congratulatory crack-hit each time the numbers went up. During recent investigations into the state or web-stats, it’s clear that though the web has been changed by RSS, web-stats have stayed the same.
Before blogs were big, the assumption web stats were built on was:
Number of unique clicks = Number of people
It was simple, and it almost worked (of course how much time needed to pass before a click was counted as ‘unique’ was a matter of debate, allowing people to inflate their numbers by shortening the interval.) It was safe to assume that the people consuming your content were visiting your site and reading it.
In the age of RSS this doesn’t hold true. E.g., according to Technorati SugarBank is the 9th most significant ‘porn’ blog (with Viviane delicately skimming my rear… mmmmmn.)
Bollocks. SugarBank’s probably not in the top 100 porn related blogs. Why the gross inflation? (Gross inflation’s a movie you’ve got to see by the way, it’s amazing what you can do with a lot of breath and a full condom).
Technorati’s counting links made to blogs, like SugarBank, that use the word ‘porn’ as a tag. Huge blogs like Fleshbot, which don’t use the porn tag, don’t appear. While counting links is a precise measure of how many other bloggers link to you, what does it tell you about how many people read your blog? It assumes that 50 links from blogs with a single reader each, is worth more than one link from a blog read by thousands. That’s incredibly dumb.
In the age of RSS, to have any idea of how ‘big’ a blog is you need to account for:
- Subscribers (how many there are, and how many read each post on average)
- Unique visitors (how many there are, measured the same way for each blog)
- Comments (per article and posted recently)
- Links in (where those links are from and how often they’re used)
- Trackbacks and Pingbacks
- Page-views
- Words read per visitor (is a blog people read in detail more important than one they skim? How can we measure words read anyway?)
- Downloads (if you’re podcasting, this is directly relevant. If you offer downloadable PDF’s, that needs measuring too.)
But no-one does this (yet?), and the reach of an individual blog is impossible to guage using the old tools.
Is SugarBank, with a couple of thousand unique visitors a day, over a hundred and fifty pages and roughly 250 subscribers reading daily - bigger or smaller than Podnography, with less than 10 pages, less than a hundred unique visitors a day, and over 500 subscribers reading daily?
I have little idea.
Given the number of blogs out there where’s a better solution going to come from? Shall we build our own? How?
Sam-
I would applaud anyone who comes out to provide some kind of more inclusive technique to determine blog statistics.
My stats boosted from 120/day to 320/day after your last Ms. Kitka post and my readers went from (at most) 12 to 24 [*Kitka bows gratefully*]. Most of my friends don’t read RSS feeds, but come to the site instead… and then there are those friends of mine who only care about pictures.
BUT, the fact of the matter is this: as long as I have your attention, what do I care about my stats…
Kitka
KitKa,
You’re a perfect example. It’s really hard to know what the numbers mean. It’s so hard to know how blogs are being used - especially while RSS is still a WTF? technology to most normal people. If I had the coding chops I’d have fixed this already. I think I know what we need - I just can’t build it.
Everything I do… I do it for you (that’s good Canadian cheese right there!)
Sam
I just bought mint[www.haveamint.com] for $30, I know it’s a bit expensive and since I do not have even a small traffic website to use it on, most would not see the point (learning from it mostly), maybe you could get some of the developers hanging around in the mint forums to build you a wordpress plugin (mint has an API)?
Good luck!
P.S. : I’ve always thought Brian Adams was british?
Jamal
For me the most important stats are from Feedburner - subscribers to my RSS.
… but I still check my server stats on a daily basis.
Jamal,
I have a very similar program in play myself. Nothing currently available does what I think is needed though. Blog stats need to account for off-site readers and then weight things to produce a number than can be compared. Thanks for the tip - your stats are cheaper than mine…
Miha,
I use Feedburner too but even that simply shows how many subscribers read your feed in the last 24hrs and not how many subscribers you have in total, or how many read in the past week or month. If your readers aren’t all daily it’s a really poor statistic.
Sam
Althought like I said I don’t like the designer’s club much (okay I don’t hate them, they the Dark Side of the Force, I am just jealous), I do believe Inman stuff (see I don’t belong to the club, they all address themselves by firstname over there!) does the thing with feedburner you’re looking for… Except maybe your total subscribers (but couldn’t you filter feeds + unique visitors? Anyway I don’t believe it’s possible to know an exact number, maybe an average, no?).
Tell you what, I’ll try to set up a few pages that it could monitor by say mid next week and I’ll give you a temporary login/pass so that you can test it (since I am not really using it at the moment)
Jamal - that sounds interesting. What I’m talking about is a synthesis of those numbers that provides a ranking, of the type Google provides and calls ‘relevancy’, but for blogs. I.e. which blogs are most relevant to readers regarding specific subjects.
The individual stats are all available - it’s just the analysis that’s lacking.
I use feedburner and still don’t know why there’s no easy view of new subscribers added each day. The churn and retention of subscribers is really useful information. In fact the average length of an ‘active’ subscription is a key metric.
If you search Technorati on the ‘Sex’ keyword, though, Fleshbot does come top.
And our site comes 15th sandwiched between Housewyfe and Lili (a very nice place to be!). But I don’t fool myself into thinking that we’re the 15th most important blog about sex, either - that would be ridiculous.
You’re right, though, traditional web stats don’t really cater for blogs. There is a Wordpress refer plugin which gives stats for page views, including RSS. But it only has daily stats and doesn’t help you on Typepad. Incidentally, when RSS is included the number of page views for our site is 1,000 more per day on average than ‘conventional’ page views shown in StatCounter and Sitemeter, so RSS is quite significant.
How useful it is to have people reading your RSS feed who might never visit your site and comment, though, is another matter. How do you judge a low-traffic site with a lot of comments against a high traffic site with few comments? If you’re relying on advertising income I guess high traffic is best. But I suspect that most people with personal blogs would prefer a small number of readers who actively comment than a high number of anonymous lurkers. And nirvana would be high traffic and lots of comments.
I confess to being a stats junkie, but whatever stats you look at it helps to remember to take them with a large pinch of salt.
Ed,
You’re seeing the same effects as me. It’s really impossible to divorce blog size from the number of RSS readers. There are many blogs I visit in person very rarely but read daily. Also - given the ability to place ads into RSS, even the advertising angle is affected. You could argue that a subscriber was a higher quality reader than someone who never chose to make the commitment to a blog…
When an integrated web stats system appears - it seems there’ll be plenty of takers…
Thanks for the nice mention. It’s more like eating your dust, instead of delicately skimming your rear. I’m a good 50 links away!
Viv you’ll surpass me. I’m writing original content so it’s way harder for me to produce as exciting material on my own than a blog which aggregates. It’s interesting to observe the differences though.
Don’t forget me when you’re (more) famous.
Sam,
You will be I think very interested by this :
http://www.adammessinger.com/2005/09/18/textdrive-gives-urchin-the-fin ger
Interesting Jamal. I wonder what they come up with…
I just registered the link to my Anemone post, thanks to a WordPress plugin I have installed. Unfortunately, the project seems to have been abandoned recently due to a lack of availability by the part-time developers and the advent of an official TextDrive-developed stats solution (as opposed to a community-developed one).
You can find more information here:
http://forum.textdrive.com/viewtopic.php?id=8360
I’ll have to update my post with the news, I guess.