Who Are The Top Biotech Influencers on Twitter?

About six months ago, Andrew Maynard (@2020Science) and Mike Chelen (@MikeChelen) analyzed the influence of the then 400 ‘Scientwists’ on Twitter compiled by David Bradley (@sciencebase). Since then, I’ve wanted to do something similar for Biotech (Biotwech?) Twitterers, but the application they used no longer works in batch.

Mike Chelen came to my rescue and created an API which will work in batch for Klout, a similar application which gives Twitter influence along with other metrics and classifications for user types. Using the list of users from the Biotech Twibe, including the top 25 influencers from the Wefollow Biotech tag, I created the bubble chart visualization using the ‘Many Eyes’ application. The size of each users’ bubble, which is directly proportional to their influence, is based on the Klout score, which ranges from 0 to 100 and is based on 25 metrics including the size of your ‘engaged’ network, retweets, and number of Twitter lists.

Top 20 Biotech Influencers on Twitter
No. twitter_screen_name kscore kclass Male/Female
1 Wise_Diva 51.11 Persona Female
2 MaverickNY 44.81 Persona Female
3 genomicslawyer 41.67 Persona Male
4 FierceBiotech 37.59 Persona Female
5 ellenhoenig 35.64 Activist Female
6 Comprendia 35.07 Activist Female
7 JohnSharp 33.48 Connector Male
8 girlscientist 32.37 Connector Female
9 salesrecruiter 32.28 Persona Female
10 Roche_com 31.35 Persona Unknown
11 2healthguru 31.24 Specialist Male
12 shefaly 31.12 Activist Female
13 sheffi 30.98 Connector Male
14 sdbn 29.94 Persona Female
15 val1a 27.64 Persona Female
16 scisu 27.53 Connector Female
17 idtdna 27.12 Connector Unknown
18 iambiotech 26.82 Connector Unknown
19 hollybarnes 26.68 Connector Female
20 CenterWatch 26.6 Curator Unknown

You can view the full Twitter Biotech Influence data set to see more details, and even interact more with the visualization as well. Those of us who are in the Twitter space will not be surprised at most who are in the list of the top twenty, but there are also some interesting observations to be made. Because Twibes are not based on the content a user tweets, there can be some spurious results. For example, while Wise_Diva is certainly influential, she rarely tweets about Biotech, but since it is one of her interests, she joined the Twibe.

Someday soon, hopefully we’ll have applications which will take tweet content into the equation, and we’ll be able to find the key influencers based on keywords (currently, on Klout, if you log in and go to the ‘dashboard’ you can search for influencers on some topics–biology and pharmacy seem to be the closest matches). Klout does list the content of a users’ tweets, but using this information directly may not accurately portend their tweet content. For example, Sally Church aka @MaverickNY is the top ‘true’ biotech influencer, but her content is listed as cancer, FDA, Google, iPhone, Boston, and social media. We do know, however, that she tweets often about Biotech, but just does not use the term. (And, would you move to Boston already Sally since it’s clear you love the place?) So, likely the content algorithms will need to be very smart, or we’ll need to do some curation. Also, you should consider that if one of your goals is to be an influencer for a certain keyword, that you should use the keyword in your tweets, as you will not always be able to count on human curation.

Another interesting trend is the fact that the top influencers are categorized by Klout as personas, activists, and connectors. While the Klout website doesn’t describe how these categorizations are made, the descriptions they give are listed in the table below. The Network Score, Amplification Score, and True Reach values were capricious, with several being zero for many accounts, so we did not analyze them (but you can visualize them on the Many Eyes site using the menu on the lower left). Klout appeared to be better than tools like Wefollow at discerning between active and inactive accounts, as the latter has at least two inactive accounts ranked highly, including @IamBiotech which is #1. However, some of the top ranking accounts, notably @sdbn at #14 which I manage, are partial bots (mine fed from RSS), and Klout seems to not discern that these accounts are not 100% ‘human.’

Klout Classifications for Most Top Biotech Influencers
Klout classification Description
Persona You have built a personal brand around your identity. There is a good chance that you work in social media or marketing but you might even be famous in real life. Being a persona is not just about having a ton of followers, to make it to the top right corner you need to engage with your audience. Make no mistake about it though, when you talk people listen.
Activist You’ve got an idea or cause you want to share with the world and you’ve found the perfect medium for it. Your audience counts on you to champion your cause.
Connector You are a constant source of information to your friends and co-workers. There is a good chance that you probably introduced several of your friends to Twitter. Your taste and opinion is respected and your judgment is trusted.

Another interesting observation is that 12 of the 16 top influencers’ ‘known’ genders are female, or 75%. Are women better at fostering microblogging networks? This finding is particularly interesting in light of the recent reports that science blogging networks have significantly more male than female bloggers. While it is tempting to speculate why, it should be noted that Twitter users are 60% female. However, it’s been shown that men are likely to get more followers.

Surely I’ve missed some biotech Tweeters, I thought about adding them separately, but my vestigial scientific tendencies told me that it would make the ‘experiment’ less reproducible (also why I left Wise_Diva in). The Twibes application is the only one I’ve found that will output a list of users–Wefollow and Twitter Lists do not. Please add yourself to the Biotech Twibe and I will redo this visualization in a month or so. I’d like also to make some tag clouds of all of the tweets, especially the influencers, but alas I can’t find a tool for that.

I’d like to give special recognition to Mike Chelen for creating the Klout API. Often I get frustrated with analyzing social media data, and the scientists who ‘get’ social media are really fantastic resources. Mike did this for us for free, but you should check out his Linkedin profile and consider hiring him for your project. The intersection of social media and science is one of the most fascinating aspects of my job, and I think exciting things are on the horizon. For a teaser, see how some scientists are using a Bioinformatics/statistics program to create word clouds of Pubmed abstracts. Of course, the Batch Klout API he developed can be used to analyze any list…have at it!

Finally, why is this study important for life science and biotech companies? Let’s say we’re able to find the top influencer for PCR, lung cancer, kinase research etc. These accounts, whether they’re individuals or brands, will be able to spread a message quickly and effectively. Of course, the issue of the ‘separation of church and state’ in media and advertising will still have a play, but we’ve talked about how this can be done carefully by companies. Indeed, Klout is already thinking about this for the larger brands, shouldn’t you be thinking about it for your life science brand?

To share this post easily, cut and paste: Who Are The Top Biotech Influencers on Twitter? http://bit.ly/bio_inf

Tagged with:

5 thoughts on “Who Are The Top Biotech Influencers on Twitter?”

  • Oh my, I most certainly would not expect to see myself on this list, never mind near the top of such an esteemed group. The first biotech people I thought of were actually @divabiotech and @shwen although neither appears here.

    Certainly we have used Klout with limited success in the past to look at Pharma companies, although I never actually looked at my own profile. We found that the analysis can be very limited by the inherent definitions of the tool, ie a top influencer of what exactly and in which specific community?

    Ultimately, we need to develop more granular and specific tools that will help focus the research more towards the users specific goals. Pharma might want to know how they are perceived by advocacy groups, patients/consumers or physicians, for example.

    As for Boston, that is definitely on the cards for next year. My biggest surprise, however, is that the important word ‘football’ did not show up :)

  • Glad to see the Klout API script worked ok! Here is the code if anyone wants to run their own copy: http://gist.github.com/590242
    It seems to return only about 20 names at once, a newer version should automatically handle larger lists by splitting them into batches.

    Interesting to see which Twibe members Klout thinks have the most influence. It is true that those at the top may have high overall influence, while relatively less with other members of the Twibe.

    Another useful feature that Klout provides is a list of people who influence a given user. If the influencers were found for every member of the Biotech Twibe it could show which are the most common. They would be certain to have a high influence within the Twibe, though it is possible that they focus on more general topics.

    PeerIndex looks promising. Because topics are discovered automatically, like Klout keywords, it is important to see how well the detected topic matches the actual content of the posts.

    Mary: There is an RSS feed of all posts in the Twibe: http://www.twibes.com/group/biotech/tweets.rss
    A script could be used to collect new posts and should have some usable data after a few days or weeks, since old posts would not be available.

Comments are closed.