Assessing Blogs
By Mat Morrison April 16th, 2007
In Blogging · Buzz & sentiment analysis · Customer experience · Stories
Has anyone got a good metric for assessing blogs? I’m thinking about something that combines (say) recency and frequency of posts, comments per post (or total comments) and inbound links. I think that should cover it. But how?
To know that, I’m going to have to set an objective or two. Watch this space.
1 Ade // Apr 16, 2007 at 10:49 pm
We’ve just been through this exercise. You’ve pretty much covered the list we used, but as we were covering some mass media blogs (e.g. the Guardian) we wanted to get a feel for the level of engagement, so highlighted any where the author responded to the public’s comments. It’s far from an exact science, but a useful pointer.
2 Mat Morrison // Apr 17, 2007 at 7:55 am
Makes sense, but - sadly for me - makes it much harder to automate! We’ve spent the last year talking to all (well, most) of the social network analysis ASPs out there, and the guys who offer Sentiment Analysis.
I’ve been wondering what layers of analysis we could place on top of this to add some value; but thinking (I suppose) all the time about an automatic tool.
I’m pleased that I’m not too far off the mark. I’ve been wondering about spam comments, too — how do I discount them? After all, a poorly-run, poorly-implemented blog will have much, much spamdexing activity.
I have a sneaking suspicion, however, that this can be fixed in the mix, by weighting the variables sensibly.
But it all depends on what I’m trying to achieve. Am I looking for:
Answers on a postcard, please.
3 Will McInnes // Apr 17, 2007 at 4:27 pm
Social Meter can be used alongside other measures but isn’t quite good enough for me to be satisfied in its use as a single measure:
http://www.socialmeter.com/
4 Mat Morrison // Apr 17, 2007 at 5:58 pm
Yeah - I rather like that. Antony Mayfield over at Spannerworks showed us a couple of hacks* that sort of feed into this. Amongst other things hey do a quick-and-dirty version of influence mapping that gives higher weight to links on Wikipedia , social bookmarking sites, crowdsourced editorial sites like Digg, and the like.
I agree - inbound links are a good way of measuring a blog. What’s good enough for Google and Technorati should probably be good enough for me.
But I need more. I need a correlation between the score, and my objective. This will let me optimise activity against goals.
I’ll give everyone a free example:
Database marketing guys often use a technique called RFM scoring (it’s like RFM segmentation, only even simpler) to predict (& control) the ROI on a mailing.
It looks something like this:
((Frequency / Tenure) x (Monetary Value / Frequency)) / abs(Latency - Recency)
On the whole, Recency is a very good indicator of “likelihood of response”, while “Tenure”, and “Monetary Value” (spend-to-date) are good indicators of potential value.
CRM/Database Marketing guys have it easy; the return they see is measurable, and readily understood.
What is the return on buzz? Or - rather - which return should I care about?
* The neatest hack was really simple: to find out what people like, do a tag-search on del.icio.us for your topic area “+cool”. So, for example:
http://del.icio.us/tag/cool+insurance
actually throws up some interesting research
5 Mat Morrison // Apr 17, 2007 at 6:22 pm
Oh — and Will; socialmeter doesn’t check for non-canonical URLs at the same time. This is a bit annoying: it means that there are two readings for every site.
http://www.nixonmcinnes.co.uk - 428
nixonmcinnes.co.uk - 140
Because this is a pretty simple scoring system (every link counts as “1″, no matter where it’s from), it’s easy to get around this. But I think we could build a better socialmeter.
6 Popularity… at Ryan Morrison & MacMillan Ltd. // May 3, 2007 at 12:20 pm
[...] This is leading to some peculiarities. I’m not convinced that Iain’s “p-books are still cool” article is, in fact, the most popular post on the site. But this could be jealousy. My reason for playing around with this comes - in part - from my interest in blog assessment. [...]