Social media and Google Analytics — who’s interested in botany?

We’re com­ing round to our annual assess­ment here at AoBBlog. Spurred by JSTOR Plants and their blog post on Social Media we’re shar­ing our thoughts. And if you have bet­ter thoughts we’d wel­come them as a com­ment on this post.

In our case we have three social net­works we tar­get: Facebook, Twitter and Google+. Facebook and Twitter are easy to pop­u­late, we have auto-posting from the web­log to Facebook and Twitter so you don’t have to keep vis­it­ing here to see what’s new. Google+ has to be pos­ted to manu­ally. We’ve also played with Reddit, Stumbleupon and Pinterest over the past year. If you’d like to see what refer­rals we’ve had to this site from social media, here’s an anim­a­tion of the past year:

Note that it’s a log­ar­ithmic scale hori­zontal scale for the num­ber of vis­it­ors. This is because there was such a spike from Reddit that, if we had a lin­ear scale, everything would appear to be stuck on the zero vis­it­ors axis, except for a quick flurry from Reddit in March. In fact you can see our year on Reddit below:

So what is hap­pen­ing here? It’s simply that without effort from us, one of posts briefly went viral on Reddit. This happened after I’d given up on Reddit. If you’re not famil­iar Reddit is a site where people sub­mit links and people vote them up, down or ignore them. I’d been sub­mit­ting some sites, not all AoB Blog posts, to Reddit. Because it’s an interest-based com­munity I’d only picked links that I thought were inter­est­ing and I think they all pretty much sank without trace. I’d been also vot­ing up other stor­ies I’d found inter­est­ing, but not many. My con­clu­sion here is that Reddit is a good source of traffic, but it needs spe­cial atten­tion and it works best when you’re fully par­ti­cip­at­ing in Reddit. I don’t have the time to devote to that. In par­tic­u­lar I think it’s best to focus on a good qual­ity sub-reddit, a page that cov­ers a spe­cific sub­ject. The qual­ity of sub-reddits vary. Some are notori­ous for being stuffed with cranks, but I think that on the whole the Biological Sciences sub-reddits, includ­ing Botany are pretty sane. They’re also unpre­dict­able. The spike you see hap­pens in March. The post that got picked up was our Valentine’s Day post on Rainbow Roses. If we had someone who could con­trib­ute to Botany as a whole, and not just our posts, this could be a net­work worth invest­ing time in.

The most favoured net­work for people who don’t do social media is Facebook. People have heard of Facebook, even if they don’t use it. We’re on Facebook at http://​www​.face​book​.com/​A​n​n​a​l​s​O​f​B​o​t​any and you’re wel­come to join us there.

I think every­one in the office now has a Facebook account of some sort, but I don’t know how many of us see it as a net­work for dis­cuss­ing pro­fes­sional mat­ters. This is why we need a mul­tiple net­work strategy. It’s not that one net­work is bet­ter than the oth­ers, it’s they they get used dif­fer­ently. For example for friends I use Facebook, for interests I use Google+, for fol­low­ing what new in blog posts I use Twitter and when I don’t want to be social in any way I use LinkedIn. But this is how I use net­works, because that’s how they work best for me. For other people it works dif­fer­ently and we shouldn’t be in the busi­ness of telling people how they should use their net­works. For Facebook we see some vari­ation in use over the course of the year. In all the graphs below the vis­itor scale is lin­ear, so the fur­ther right a dot goes the more vis­it­ors the blog is get­ting from a site. The ver­tical axis is pages viewed per visit, so the higher it goes the more of the site a vis­itor is brows­ing. Each dot on the trail marks the end of a month and we start on June 1:

What we see here is that there’s a very dif­fer­ent use in September to November, February and March. In September what we had was a rel­at­ively high num­ber of page views per vis­itor. Initially I thought this was high interest/engagement in the site but that might be a bad assump­tion. What we seem to see is people at the start of term look­ing around to see what the site offers and then hit­ting the site in num­bers dur­ing term time. Now is this higher view­ing, but lower pages per view sig­nal with the spikes to the low right really a case of more vis­its and less engage­ment? Another pos­sib­il­ity isn’t that people are less engaged, but are more famil­iar with the site, so they’re hit­ting what they want and mov­ing on to the journal rather than wan­der­ing around try­ing to find some­thing rel­ev­ant. But it also might not be. What the graph doesn’t show is that we had a massive uptake in likes dur­ing October. A lot of the vis­it­ors for this period must be people who weren’t famil­i­ar­ising them­selves with the blog in September. Taking this into account, it looks like the Facebook vis­itor is someone who has a spe­cific tar­get they’re inter­ested in when they visit.

We prob­ably see a sim­ilar pat­tern for Twitter. In the graph below the dots are all sites start­ing with ‘T’. This was the only way to get refer­rals from Twitter​.com and on to the same graph. During the past year Twitter phased out twit​ter​.com and made it’s refer­ral ser­vice with short links. In the next year should account for all Twitter referrals:

Posting to Twitter is a zero-effort task. Autoposting plu­gins mean that blog posts are tweeted when they’re pub­lished. The rel­at­ively low num­ber of vis­its from Twitter is prob­ably partly a reflec­tion of this lack of effort. We did innov­ate with Twitter by cov­er­ing tweets at IBC18 from a dis­tance, which was a high-effort under­tak­ing. There was pos­it­ive feed­back for this, and tweet­ing from other con­fer­ences through the annbot account would increase interest in the twit­ter feed, if not clicks-through. This does require people at con­fer­ences who are will­ing to tweet what is happening.

The new launch over the past year has been Google+. Take-up of Google+ has been patchy and user’s opin­ions have been mixed. A com­mon com­plaint is that Google+ is a ghost town and no one posts there. This isn’t _entirely_ accur­ate. Google+ has pri­vacy set­tings and com­monly people only share with other people in their circles. A new mem­ber of Google+ isn’t likely to be in anyone’s circles and so won’t see much of the con­ver­sa­tion that’s hap­pen­ing. All AoB edit­or­ial staff on Google+ reg­u­larly post items with Public vis­ib­il­ity, so new mem­bers can see what is being pos­ted. The Annals of Botany page on Google+ also routinely posts pub­licly. We have dis­cussed set­ting up spe­cific circles to post extra con­tent to for fol­low­ers, but so far have pre­ferred to leave everything as pub­lic as pos­sible. Tracking Google+ vis­its is dif­fi­cult. plus​.google​.com and plus​.url​.google​.com are obvi­ous refer­rers. Is google​.com/ refer­ral also a refer­ral from Google+, or does this cover Gmail and Google and not Google+? All three make an appear­ance in the graph below:

Here the main story is that the Google+ fol­low­ing is grow­ing. Whether our use of Google+ is improv­ing or get­ting worse is swamped by this sig­nal. There is a second factor. Google+ vis­it­ors seem to aver­age more page views that Facebook vis­it­ors. If we equate page views with engage­ment then it would sug­gest that the aver­age Google+ vis­itor is enga­ging more with the web­log. This plays into the con­trast of Google+ as an interest net­work versus Facebook as a social net­work, but it’s prob­ably too early to say this is a defin­it­ive dif­fer­ence. The next year will provide more data. Another fea­ture we have found with Google+ is that the Annals of Botany page here already has more fol­low­ers than the equi­val­ent page on Facebook. This then raises the ques­tion why is Facebook pro­du­cing more refer­rals? Is this a fail­ure in how we inter­act on Google+. This might in fact be a ques­tion of demographics.

On Facebook we have tar­geted under­gradu­ates. These are people study­ing all things botan­ical and so much of what we post is rel­ev­ant. On Google+ people are sign­ing up through interest. In the case of pro­fes­sional research­ers these are people with spe­cific interests. A bioin­form­atician isn’t neces­sar­ily going to want to read an eth­no­botan­ical study of the domest­ic­a­tion of cacti. Oddly for a journal we don’t simply need to be vis­ible, we also need to be ignor­able.

People use social media as a fil­ter for inform­a­tion. Realistically we have to accept not every paper in Annals of Botany or AoB PLANTS will be vital read­ing for every­one. Acting as if it is makes us anti­so­cial. An example is the publisher’s mail­ing list you signed up for and now can’t fig­ure out how to unsub­scribe from. What hap­pens is that the emails get spam-binned. The pub­lisher con­tin­ues in bliss send­ing out thou­sands of emails, many of which will never be read. In the case of social media the spam bin doesn’t exist but it’s easy to unlike or unfol­low. Dominating people’s social feeds with every detail will make us emin­ently unfollowable.

Rather than get­ting around people’s fil­ters the smart move is to be part of the fil­ter. That means con­trib­ut­ing to peoples exper­i­ence and shar­ing mater­ial that isn’t Annals-originated. On Facebook and Twitter we do with via the Scoop​.It sys­tem. On Google+ we post mater­ial manu­ally. If we post a link to an AoB paper that a lot of people fol­low that’s a meas­ur­able suc­cess. However if we post a link to a PLoS One paper that a lot of people fol­low, that’s a suc­cess too, because we’re being use­ful and help­ful con­trib­ut­ors to the Botany com­munity. It is how­ever often a pain to meas­ure how suc­cess­ful links to other sites are, because we can’t track their arrivals. The rel­at­ively high num­ber of Google+ fol­low­ers sug­gests that while we aren’t get­ting the clicks-through that we do from Facebook, we are achiev­ing our aim of being more of a help than a pest for research­ers and the public.

Alun Salt. ORCID 0000-0002-1261-4283

When he's not the web developer for AoB Blog, Alun Salt researches something that could be mistaken for the archaeology of science. His current research is about whether there's such a thing as scientific heritage and if there is how would you recognise it?