What do all these numbers mean?

Bloggers, or anyone who maintains an online profile, have an ambiguous relationship with visitor stats and data. On the one hand we like to dismiss them as meaningless, but then secretly feel chuffed when we can outscore someone. I've tried to promote them as one way of measuring impact, but with the caveat that context is important. For instance, if you're a blogger in a relatively obscure area, such as Barry Town football club, then your range is limited and unlikely to compare in absolute numbers with, say, a blog reviewing Apple products.

I recently passed 300,000 views on this blog, over about 700 posts - that's not as exciting as it might sound as I've been going since about 2006. My friend Liam says he gets about 200K a year on his mobile tech blog. Here are some more stats from tools I use:

Slideshare - 220,000 views over 6 years on 59 presentations
Blipfoto - 92,000 views over 420 entries
Flickr - 49,000 views over 1,140 photos
Citations on academic articles - 1,620 (according to Google)

The first thing to ask is how reliable is this data? Tony can probably answer that better than me, but I think the blog traffic doesn't take into people who subscribe via RSS, so in that case underestimates. At the same time it does include my own obsessive self-clicking, so that may even out. It may also include bot traffic which doesn't really mean anything at all.

But if we take the data as accurate for now, the more interesting question is what does it mean? I would like to make the case that 300,000 visits to my blog equals 300,000 careful, considered reading of my posts, but the data doesn't back that up. Google analytics tells me that my average visit duration is 57 seconds. I suppose that is just about the time it takes to read a blog post, but it probably doesn't compare well with time taken on reading academic articles. My bounce rate is high too, at around 85%, so people come in, read a post and then go elsewhere, but I think that's fairly typical for blogs, I'm not providing a newspaper where you want people to go from one feature to another.

Blipfoto probably has the highest return on audience per post, and does better than Flickr. This provides a nice comparison as I have linked my blipfoto account to twitter, so it automatically tweets when I post, but I haven't done this for Flickr (which I think of more as a repository). I get 219 views per post on blipfoto versus 42 on Flickr, so we could say that the effect of linking to twitter is a 5-fold increase in traffic (if that's what you want). But I would say that Blipfoto is the most personal of these tools and the least connected to my professional life, so I wouldn't argue for any professional impact here.

What I think we're struggling with here is the newness of all this. What is the value of a slideshare view versus a download of a paper? Is there a conversion rate? I've tried to keep a foot in both the traditional camp, through publications, and the new forms of dissemination, so these figures might offer a reasonable comparison between the two. Looking at my citations you could crudely say that one citation is worth about 200 blog views (or vice versa).

However, it's probably futile to try and measure these things. It's not as if citations are a particularly reliable metric anyway, so we shouldn't use those as a currency to convert everything else into. It will also vary widely depending on discipline, and individual. And more importantly, we already have plenty of systems that our career path requires us to game, so why invent more?

Make no mistake, turning altmetrics into performance measures linked to real things like promotion, job security, money would have negative effects. For example, there is a correlation between number of posts and views, so if I was purely chasing the numbers I should post more often, regardless of quality (what quality?), or I might think I could get better numbers from splitting this post into three separate ones. This already happens in traditional outputs, I remember someone advising me not to write books because for the same effort I could have 5 REFable publications. This return on investment view ignores the internal validity of the form. I also tend to think of blogging as a social function, connecting with friends and peers, but if I'm linking it to direct returns then it's another example of monetising relationships that Brian is so uneasy about.

But when I've tried to make the case for digital scholarship and tenure, it is something along these lines that I want to argue - that the impact you see here is as valid as the impact we have chosen to recognise and measure. So we need to start thinking about what these numbers do mean for academics - we just need to make sure we don't turn them into the thing itself, with all of us chasing and parading numbers. But maybe that's inevitable as soon as you shine a light on them, in which case, ignore this post.