How does blog readership grow, and (how) does this matter?
How useful are science blogs, and what is the future of science blogging? Probably the worst place to get a useful answer would be from a blog. But on the other hand, it’s the blogs that have relevant data, and it’s bloggers that think about this stuff. Here are my opinions and some digits, and my inferences.
For recent context: Jeff Ollerton has mentioned how his blog has slowly but steadily grown. (After a lag phase, it appears growth has been linear. At least, that’s my bet for the most parsimonious model). Jeremy Fox posted a lot of careful thoughts about whether the ecology blogging community is saturated, empty, or none of the above. The discussion in the comments of Fox’s post is really interesting, about who, why, where, how and when blogging happens in our field. Paige Brown Jarreau, whose dissertation is about science blogs, responded to Fox’s post with oodles of data and links, and that’s an interesting read if a post like this one is also interesting to you.
I’m convinced we are still in the early days of academic blogs, at least in biology.
First of all, there are many professional blogs, but few that are updated frequently, and few that are widely read. A new blog can accumulate an audience as long as posts are interesting and frequent. Take, for example, Scientist Sees Squirrel by Stephen Heard. He has interesting and insightful ideas that are worth our time, and just a couple months in, he’s seeing positive effects.
A lot of scientists have very different — and important, interesting, and useful — things to say. If they chose to say them regularly in a blog, and say them well, then the result would be a well-read blog. It might not take a mountain of work, but it does require a consistent application of effort. (The ability to allocate that time is unevenly distributed in academia — this is discussed really well within the comments on the Dynamic Ecology post by Fox.)
So, there are multiple oodles of room for anybody to step in and run a quality blog. It’s just that few people are doing so. It’s not a small investment to make. The way that blogs get a readership in the scientific community is very different than the peer-reviewed papers. When we publish a paper, we expect others who work on similar questions to find and read our papers, and cite them as appropriate. A good paper, in a well-read journal, about a contemporary issue, has a good chance of being well-read and well-cited, even if the authors are unknown.
On the other hand, a high quality blog post on a new blog isn’t going to have much visibility and there’s a good chance that the target audience might not see it. Blog posts aren’t indexed* and the scientific community isn’t trained or expected to read blog posts. On the other hand, if you take the same post, and put it on a blog that has a steady readership, it is guaranteed to have more visibility.
What’s the difference between a high visibility blog post and low visibility blog post?
Time.
There are other variables that matter, but time appears to overwhelm all of the others. There are wonderful posts that I wrote a couple years ago that almost nobody sees. There are mediocre posts I’ve recently written that are being seen way more than my amazing posts from a couple years ago. That’s just the nature of blog audiences.
People say that for a blog to be successful, you’ve got to stick with it. I wanted to evaluate the effect of time on readership, so I ran some numbers.
I don’t have analytics on this site, just some weak statistics provided by the hosting site, wordpress. The most common currency that people use is views, so I’ll go with that. (This doesn’t include people who get email subscriptions and read by RSS, but it’ll do.)
Here’s how views per month** have changed over the 2.1 years this site has existed:
It looks like growth is sporadic, and has slowed down, or flatlined. But I remembered that we used to post very frequently, nearly every weekday. Now the site is down to about two per week. Because new posts bring more visits, I then adjusted to consider views per month per post:
And I thought that relationship was really striking. There are three clear outliers. The huge positive one is caused by a single post that caught a lot of attention (in which I used a chimeric epithet to bring attention to the sexist behavior of a certain space scientist). The two negative outliers are the two Januaries in the dataset. Januaries are, apparently, not good for blog readership, with people tied up with holidays, travel, the new semester, and so on. If we toss out three outliers, here’s what we get:
That relationship has an r2 of 0.9.
As an ecologist I’m not used to such tight relationships. Obviously, at least for this site, time-since-start-of-blog is the only real predictor of how many views a post will get. That’s encouraging for the future, but discouraging if you think a single post is going to make a site grow.
Last month (February 2015) wasn’t the best month for the site, in terms of what we had to say. It wasn’t a bad one, but it wasn’t special by any means. Nonetheless, it was better than the other months. Why? Because it’s the most recent month. I like the positive trajectory, and have no idea how much longer it’ll persist. But boy-o, it’s weird to think that the only thing that affects audience is time.
Of course what we say on the site matters. If this relationship holds true for other sites (and I have no idea if it does, though I’d love to find out), then perhaps what individuals have control over is the slope of the relationship. You would think that good blogs grow more quickly, and not-so-good ones won’t grow as quickly. That is, if “good” means, “generates readership,” which is not a concept I buy into.
Jeremy Fox wrote, “I’ve decided that in blogging it’s the shooting match, not the individual shot, that counts. No single post matters that much in the long run. But a blog as a whole can have a large cumulative influence over time.”
I wasn’t so sure I agreed with that point, but then when I look at the numbers, I’ve definitely come to this line of thought. No matter what we do here, the audience only responds to the site over the arc of time. And so whatever effect there may be is also exerted over the arc of time. Of course, I’m aware that other bloggers who might read this will chuckle, considering that two years is just a blip compared to what others have been doing. That’s both fair and spot-on. I have no idea what’ll happen over the next several years, and if the relationship remains linear, that’d be mighty amazing to me.
An alternative interpretation is that the linear growth of readership is not a bug, but a feature. Instead of a constraint, it could be seen as a consequence of good things happening here. To some extent that’s got to be true, because the slope is above zero. But it can’t explain how the fact that the points fit the line so well, when I qualitatively know that some months have been waaaay better than other months. It could indicate that every month, we’ve been equally effective every month at doing good things to build readership. That’s an overly charitable interpretation, I think. I think any individual post, or any individual month, can’t do much to build readership. It just takes time. Obviously.
But what about that positive outlier? What can I do to break away from the time constraint on readership? I got lots of views from a single post, but that post wasn’t helpful for the blog, or for me, or for the world. I knocked it out very quickly in response to a current event right after it happened, and those who took a little more time to reflect and write carefully were more constructive. So if the only thing that I can do to break the curve is to push the margins of respectability, then I’m not really inclined to give it another shot. I got a lot of views from being bombastic and a little rude. That’s not a good lesson, and it’s not a good model for the site, even if it would increase views***.
If you look at how blogs grow — or at least how they can grow with proper sunlight and nutrients, then the decision to set one up might make more sense, maybe? If blogs can grow at a linear rate, then I would bet that the slope of the blog would have to do with the size of the target audience and the quality of the blog (whatever that is).
I hope you see this as an argument for starting a blog, provided that you are interested in investing the time, and I hope that this is also an argument that there is plenty of space for growth in academic blogs. Just like a mortgage, or an academic career in general, it takes a steady investment to build capacity. But considering how few academic blogs there are right now that people read on a regular basis, compared to how many people might choose to read, then the sooner you start, the sooner your audience will grow, provided that you feed and water the blog regularly.
Is blogging a moribund medium? Um, not for me, and I ain’t more special than all of you.
—
*The folks at The Winnower want to index blog posts, for what it’s worth. No, I’m not about to put a DOI on every one of my blog posts, though if someone wants to do that, go for it.
** I struggled with whether to provide the units on the y axis for these figures. I think there are some good reasons to not give this information. But if this post is all about building audiences, then I figure a lot of people would be really wondering about the units, so I might as well just put it out there. To some, they are small numbers, and to others, they’re huge. From my perspective, it’s really huge and I’m still trying to get my head around it.
***I think it’s no accident that some of the most popular blogs by scientists have that kind of predilection, and I’m making a point of avoiding such a route. It might cut r, but should also increase k.