Thoughts on the unpredictability of which social media posts will be popular

I tweet a lot and I write a lot on my blogs, especially on my ACGT blog. Sometimes I write things where I think: This…this is the one that will go viral. And of course it never happens that way. The more I try to engineer a meme, the less likely the chance of success seems to be. This is why my concept of a #MexicanTweetOff never went anywhere, and that no-one seemed to get the joke in my recent #IamSpartacash tweet:

But then there are those tweets and blog posts which I think won't be of that much interest to anyone, but turn out to be the most popular things I've ever written. This week has been surprising in that two of my social media posts have taken on a life of their own.

Everyone loves a good gnome story

First, I spotted the opportunity to have a bit of fun with Genomics England. Last Thursday was a big news day for them with the project being featured in Nature. So after seeing the piece in Nature early on Thursday morning, I came up with this:

This tweet went on to have over 40 retweets and Genomics England ended up featuring me in a storify article on Friday about their news of the previous day. To date, that tweet has reached an audience of over 67,000 people on twitter!

Microsoft as a tool for bioinformatics

Then on Friday I was trying to find an image to use in a talk. I wanted to make a point about Excel often being an inappropriate tool for the management/querying of biological data, and just wanted a picture of Excel containing some sort of biological data. I was quite surprised by the image that I eventually found, and immediately shared it in a quick blog post:

Admittedly. the 'NFSW' part of the blog post title was a tactic designed to deliberately provoke curiosity. I tweeted about the blog post on Friday afternoon, a time which normally doesn't provoke a lot of interest (many of my European followers will have gone to bed by this point).

But it seems that this post hit a nerve and it has subsequently taken on a life of its own. Currently, there has been 64 retweets and 44 favorites, and now it is being spread on Facebook. By the end of Friday, it had become the most read item on my blog for the entire week (this never happens for posts I publish on Friday afternoons). I assumed that traffic would die down on Saturday but that didn't happen. By the end of Saturday, the post was my most read article of anything I have posted in the last 4 months.

At this point I assumed that things would surely quiet down on Sunday, but that didn't happen either. Traffic to my blog doubled compared to Saturday, and the post has now become the most viewed article of anything I have written in 2015, with almost 3x the page views compared to the next most-read article.

I guess the message here is that I should stop trying to predict the popularity of my social media posts!

No more science…at least not on this blog

I've decided that as I'm increasingly using this blog to write about science-related topics (bioinformatics in particular), that they should have a new home. Therefore, all of my existing genomics and bioinformatics blog posts have been copied to my new ACGT blog:

I have also removed these posts from this site so that (ultimately) they will only show up in one place in search engine results. Going forward, I will use the ACGT blog to write about anything that is predominantly science related. In contrast, this blog will now be used almost exclusively for stuff that is not primarily related to my research work.


Updated 2015-11-12  to reflect that I have now removed the posts that were duplicated between this blog and ACGT.

7 reasons why I don't like content 'aggregators' who scrape blog sites

Today a post on twitter drew my attention to Bioinfo-Bloggers, a site that aggregates content — i.e. the full blog post is reproduced — from 28 different bloggers who write about bioinformatics and genomics.

Outwardly, this might seem like a good idea. The bloggers get more exposure to their material, and readers can visit just one site instead of 28 separate RSS feeds. However, there are several reasons why I have issues with this type of aggregation. Many of my concerns apply even when individual bloggers have expressly licensed their material for reuse (e.g. by use of a CC0 Creative Commons license).

  1. The site lists the 28 blogs as 'contributors' and lists the blog writers as 'authors'. This strongly suggests that the people in question have consented to their material being used, even when this is not the case.
  2. Links to the original blog posts are included, but only at the end of each reproduced entry. The included text says that 'This is a syndicated post', further suggesting that the original authors agreed to have their content syndicated.
  3. The Bioinfo-Bloggers website asserts copyright over all material (see footer section of website).
  4. The original bloggers lose web traffic. This can matter for minor reasons such as when you want to include details of how popular your blog is for outreach sections on research grants. But it potentially — depending on how much traffic Bioinfo-bloggers gets — deprives you of knowing who is looking at your content, which articles are more popular, etc.
  5. People don't a chance to comment on your blog (unless they follow the links). You may lose some direct engagement with your readers.
  6. If people start using this site rather than viewing your blog, what happens if Bioinfo-Bloggers stops including your blog site, or shuts down altogether? In the former case, people might just assume you are not posting any more.
  7. What happens if Bioinfo-Bloggers starts including content from other blogs that you don't approve of? Your blog post may appear alongside another which espouses views you find offensive.

The first three points could easily be addressed by removing the claim of copyright over all material, by making it explicit that this site is just scraping other sites and that the original bloggers may not be aware of this, and by placing links to the original blog content at the top (not bottom) of each article.

There are currently some ongoing discussions about this on Twitter. E.g.

A new home

Moving home is listed among the most stressful events that one can experience in a lifetime. I wonder if 'moving blogs' should be included in that list too?

My first forays into blogging were around the time that Apple had just launched a set of tools under the '.Mac' umbrella. These tools were a successor to an earlier suite called 'iTools' and were in turn succeeded by the ill-fated MobileMe service (Dan Moren provides a good overview of these early transitions).

Under iTools and .Mac, Apple provided a web-based tool called HomePage that allowed you to create simple websites. As I recall, it was painfully slow to use and not very flexible. In 2006, things improved somewhat as Apple released iWeb, a desktop application that was part of the iLife suite of tools and which let you create websites that could be hosted on your iDisk (though you could also publish to other hosts).

The iWeb application made it relatively easy to create various different styles of website, and included a dedicated template for blogs. I used iWeb for a long time — hosting several different blogs, and my own personal website — before realizing that there were many things you couldn't easily do with the service. Support for iDisk-hosted websites continued as Apple transitioned to MobileMe, but the life-support cable for this service is about to yanked out, with Apple wanting people to migrate to its free — but iWeb-incompatible — iCloud service. Support for any iDisk-hosted sites will expire on June 30th and presumably these sites — along with any other files you may still have on your iDisk — will simply disappear.

Even before Apple announced iCloud, I had already been looking around for a alternative service to use for simple blogging and had started to use Posterous. Late last year, Posterous undertook some rebranding and became 'Posterous Spaces' with more of an emphasis on sharing and becoming a social network in its own right. This concerned me a little bit as I felt that Posterous's strength was in its simplicity, and Posterous Spaces seemed to suffer from a lack of focus. Further alarm bells rang earlier this year Posterous announced that they had been bought by Twitter. Although they claim that the service will "remain up and running", I'm now concerned that I can't count on Posterous as a long-term blogging solution.

All of which brings me — the writer — and you — the reader — to Squarespace. I've been hearing a lot about this company from many of the 5by5 podcasts that I listen to. For a low monthly fee, Squarespace will let you build beautiful websites to meet any need you have. So I'm now using Squarespace to host both my main website and my blog. It's been a useful opportunity to remove a lot of old content and try to give my website a much cleaner look. I'm still debating whether I should let all of my earlier blog posts remain on Posterous, or whether I should try to migrate them all to this site. In any case I'm really happy with the Squarespace service and hope to stay here for many years to come.