The topic of blog influence has been occupying my mind recently.
This is possibly the longest post in blog history. If you’re feeling fit and aggressive then do a few star jumps and shadow boxing, and dive into it. If, like me, you’re pale and wan and prone to fainting, then just have a look at it from time to time and if you start to panic, run away like a tiny white rabbit.
Influence vs popularity
So, to influence. I’ve travelled a fair distance down this road but I don’t think I’m at the end yet. At the very first step I discovered that influence is not necessarily the same as popularity.
Imagine I’m interested in financial matters. In this case – to take a UK-centric paper-media analogy – I would subscribe to the Financial Times because it’s highly influential. Heat or GQ, or even Viz, might have bigger circulations and subscription rates, but they’re not as influential. Strange how something so obvious in ‘real life’ needs pointing out for ‘virtual life’ (or R vs VR).
A bit more research shows that generally however, the two do tally. Now given that my analysis isn’t going to be totally scientific, and granted that none ever really could be, I’m prepared to run with this assumption. I’m prepared to believe that a popular blog, such as scobleizer, is also influential.
What am I trying to achieve?
So now that I’ve made my first, possibly erroneous assumption, what exactly is my objective here? Well, ideally I’d like to say ‘given the popularity (and by assumption influence) of this blog, I can give a ballpark figure and say that X number of people read it on a daily basis.’
You can spot another assumption coming along can’t you? Yes, it’s actually impossible to say with any conviction how many people read a blog. My good friend Seamus McCauley over at Virtual Economics points out that page views is a bad figure. You could have a blog that many people impulsively subscribe to, but then hardly ever read. Or you might have one that people tend to link to and even read a lot, but not subscribe to. There are also complications with ‘unusual’ uses of feeds in which people resyndicate a feed out – several examples of which you can see to the right of this very blog, as my copywriting/ PR/ journalism/ tech feeds which I syndicate out from Google Reader.
But I remain convinced that I should be able to say ‘if this blog has this many page views and this popularity rating, then by reverse calculation, this blog with this popularity rating must have X page views and is therefore Y times as influential.’
So now it’s two assumptions. But I think it’s a fair one to equate popularity with influence, and to say that, as a general rule, if I get, say, 100 views a day (I don’t – post-edit: but I do now), then it must be possible by using my popularity score to figure out, roughly, what for example engadget or gizmodo get a day. Even if we’re talking within certain percentage points of error here, at least I could give a range.
But you try looking for page views online. No one seems to want to give them out.
Oooh, it’s getting complicated
Let’s forget absolute page views for now. Let’s concentrate on a ‘popularity’ score. Here, we still have problems. It seems that no one wants to give absolute figures for this either. There are several online resources we can use to conjure up some arbitrary score, and my spreadsheet currently contains the following (with links to FG figures where possible so you can laugh at my pitiful ratings):
- Technorati- lists an ‘authority’ figure, a rank, and a links figure. The Technorati Authority is the number of blogs linking to a website in the last six months. The higher the number, the more Technorati Authority the blog has. The rank is how high (or low) you are in the Technorati order of things. The number of links shows, well, actually I’m not sure, but generally it’s higher than the authority number, so perhaps it’s the total number of links in. The problem with these figures, in particular the rank, is that there’s no way to make them ‘absolute’. I cannot walk up to someone and with any credibility say “well, I’ve done a lot of work on this and you should opt for blog X because it has a Technorati authority of Y.” I’m convinced that something more is needed. I know that people need to visualise what the authority really is.
- Blogpulse – seems to me to be a similar offering. It’s essentially a blog search engine, so you type in the blog address and you get a number of links to that address. This is handy I guess, because it’s essentially a citation. Isn’t it?
- Bloglines- is an all-in-one online blogging and aggregating solution, but also seems to offer some useful metrics, such as number of posts, feeds and citations. Remember, I’m doing a popularity search, on the basis that influence is roughly equatable to popularity, so these could be useful too.
- Alexa – is where things start to get interesting. It shows traffic figures for a website, including reach, page views and rank. Reach measures the number of users, page views is the number of pages read by Alexa toolbar users (you can see what’s coming up can’t you?), and the rank is derived from the other two. Alexa is tantalisingly close to what I need, but there are two problems. Firstly, it just expresses reach as a percentage of the total Alexa population. So it’s nice to know that Yahoo accounts for 25% of the Alexa population, but what is the total population? Just give me that number and I have the figure I so desperately yearn. But it’s nowhere to be found, presumably for good reason. And the reason is probably the second problem with these figures: it only measures a subset of the population, that is, people with Alexa toolbars. OK, so any set of figures is necessarily going to be a subset, but straight away one has to question the skew in these figures. Namely, that while my computer-consultant cousin might have it installed, my technophobic aunt almost definitely does not. Which is also probably why this blog doesn’t appear in Alexa’s ratings, and why I had to link to it using Yahoo’s URL instead.
- What else is there? Oh yes – Google and Google Blog search. Curiously no one seems to cite these as valuable research tools and yet I can see that both could be useful simply as a measure of mentions ‘out there’, especially when combined with the link: and related: parameters. One problem this throws up however is that of links from blogrolls, internal links (ie links to one’s own blog), and feed links, particularly WordPress feeds in the case of this blog. The waters are muddied further.
Imagine a spreadsheet with these figures on them for the blogs I want to analyse (with the exception of the Technorati rank and Alexa rank because, unlike the other figures, they get smaller as a blog is more popular and I don’t know how to ‘reverse’ them). That’s what I’ve got. It’s very pretty but, as you’re about to see, largely useless.
Surely it’s a simple case of adding these figures up and finding out which is the most popular, right? Well, sort of. Kind of. Maybe. First I’d like to know more about these figures and see whether they tally. Let’s find relationships between them, shall we? And this is where things become very frustrating indeed.
As far as I can tell, there is very little relationship between any of these figures. Try comparing, say, the Technorati authority with the Blogpulse messages across blogs. Their ratios will vary, widely and wildly. Same with all the others. I cannot find even two sets of figures that agree. Even the Technorati authority and links figures don’t work consistently.
This implies that there are different kinds of blogs. And certainly I set out thinking this could be the case, as I explained with the difficulty of page views. Perhaps ‘small’ blogs behave differently from ‘big’ ones. Perhaps the relationships aren’t linear.
But this still doesn’t help me much. If it’s not as simple as addition then perhaps it’s as less simple as multiplication and division. I’m after a formula here, I think.
The Holy Grail
Enter bloginfluence.net. Holy cow, it seems to be what I’m after. It is almost as if someone like me has looked around what’s available online but, unlike me, has a brain and knows what to do with it. And there’s the calculation:
Type in your blog address and out it pops. Beautiful. Other people seem to think it works too, it’s not just numbnuts like me. This could be what I’m after.
It’s only grail-shaped
But wait. There’s just one problem: it seems to have an error in the bloglines element, consistently and for any address. I’ve emailed the author about this so let’s see what happens. Meanwhile I could pop my own figures into my own spreadsheet, but I’m not absolutely sure which figures are used here from Bloglines. I imagine citations, but I could be wrong…?
Getting someone else in
If you’ve managed to read this far down this HUGE posting, you’ll have realised that this isn’t as simple and easy as I thought it might be. I mean, I knew I was making assumptions and quite possibly trying to do something impossible, but I’m nothing if not bone-headed. At least I’ve kind of proven to myself that it’s a hard thing to do. Which is why I might get someone else to do it for me.
Onalytica – which I came across via Shel Holtz’s fabulous podcast – do it all. They take the very valid position that, in order to measure influence, you need to figure out who your readers are. This is precisely the definition I mentioned right at the very beginning. But these guys really do analyse it, for you, for a price.
My question, Spinal Tap-style, would be where do you stop with this? To measure the influence of your readers then surely you have to measure the influence of their readers, and their readers’ readers, and so on? I guess one degree of readership is sufficient.
What happens now?
So perhaps Onalytica is the way forward. But hang on. Wait a minute. If they simply analyse these figures for readers, then surely I could use Technorati to analyse linked blogs too? And if I could get the bloginfluence figures to work properly, then that would give a good indication of popularity, wouldn’t it? Then I would have my popularity ratings. And, by implication, influence.
Time to get the spreadsheet out again…
I’m perfectly prepared to believe that this is something that either cannot be discovered, or that needs a much more ‘human’ oriented approach, that is by actually reading the blog(s) involved (which I have already), by monitoring linked blogs, and by getting my hands dirty.
The analogy here is with share prices perhaps, in that a successful trader doesn’t just sit by a screen, he talks to people and keeps his ear to the ground. But I do know that the reverse can be true: that the share price effectively arbitrates for all these variables and comes out at a single value reflecting a company’s current and – ideally – future performance. I also know I’m not dealing with a market here. So many assumptions…