On Monday 7 December we ran a half-day workshop as part of the POET (Public Outreach Engagement Tool) project. For this workshop we invited Twitter using academics, students and administrative staff (e.g. University communications unit; Library support) to join us for some brainstorming sessions around the features that a tool should have in order to help to optimise the effectiveness of public engagement through Twitter. The outcomes of this workshop will feed directly into the design of a Twitter tool that is currently being developed at Horizon Digital Economy Research.
Even though many workshop participants did identify specific target audiences that they have in mind when using Twitter, that audiences varied widely across the participants, raging from lay people and industry, to policy makers, academics, scientists, social sciences, interested publics, students and potential future students, distance learners, conference attendees and those that couldn’t make it to the conference
Below I summarize the results from the workshop, grouped into four main categories: Definition of ‘success’ with Twitter; Type of desired information; Presentation/ visualization of the information; General suggestions.
Definitions of Twitter ‘success’
- Increasing network within a target audience
- Being contacted by by someone who could be of interest/use in the future
- Promoting of publications (increased downloads of publication from open access repository)
- Promoting of events (increased sign-up for the event)
- Increased traffic at a linked site (downloads of publication from open access repository; sign-up for an event; visits to an information page
- Getting re-tweeted/favoured (indication that other people appreciated the content of a tweet), e.g. conference highlights
- Getting mainstream media platforms to pick up a twitter campaign
Type of desired information
- Finding a specific audience:
- Search for handle of people in related/targeted field
- Differentiate between internal (part of the University) and external audiences
- Identify the ‘type’ of twitter followers:
- Academics, public, organization
- Age, interests, who they follow
- ‘Dormant’ followers vs. frequent re-tweeters
- Most active followers and how much input they have on our twitter account
- Finding relevant ‘twitter conversations’:
- Comparison with similar users/groups
- How frequently is a particular #hashtag used?
- ‘What is trending’, but filtered to relevant #hashtags, spatial locations or target audience
- ‘Who’ is enjoying who’s tweets now
- Notification of best time to tweet:
- notification of occurrence of relevant ‘twitter conversations’
- based on activity of twitter followers
- Impact of tweets:
- Importance (number of follower) of the people who responded (re-tweeted/favoured/replied to) a tweet
- Correlating with hits to URL link in the tweet
- Correlation with followers gained and lost
- Number of people who clicked the URL link in a tweet, but did not re-tweet/favour it
- Chronological analysis -> what happened/when after a tweet? What triggered negative reponses – blocked/un-followed
- Impact of photos or images in terms of RT or favouring
- Total number of followers reached, including retweets/favouring/responding/commenting on responses …
- Identify re-tweets that have been done by followers outside of the tweeter’s country to measure international reach
- ‘Quality’ of retweet, who + how -> outside my usual ‘zone of influence’
Presentation/visualization of information
- Information about a Tweet:
- ‘Social graph’ type visualization focused on a Tweet (network of re-tweets/favouring; number of followers of re-tweeters)
- Visualize overlap of people who retweeted/favoured/mentioned different tweets (e.g. tweet A and tweet B were re-tweeted by 70% same people)
- Graphical representation of re-tweet chain + total number of followers reached.
- Colour coding for engagement rate/value/success of tweet, e.g. colour mixing for amount of replies (red) and re-tweets (green) – the more of each the greater the colour intensity: transparent to opaque
- Colour coding of tweets based on tweet milestones, e.g. 5RT = 1 colour, 10RT = another colour
- Map indicating where lots of tweets/re-tweets are, e.g. Malaysia campus – within specified time window.
- Map showing places of engagement over a certain time period not just individual tweets, e.g. has the Malaysian campus engaged?
- Information about followers:
- LinkedIn style information about 1st, 2nd etc. level connections that the follower has
- How did they find you? What did they ‘search for’ to bring them to you?
- Indication of amount of shared followers/followees
- Influencer rating -> visual/colour or some other way
- Influencer analysis -> like tumblr, reblog tree (see union metrics for details)
- Indication of users who have re-tweeted or favoured a Tweet from you for the first-time (graphical by symbol or colour?)
- Graph showing clustering of followers, influencers, etc. based on various criteria (geographic location, common followers, type of work/expertise, etc)
- General suggestions
- Graphics that use data to tell a story visually.
- Big picture/detailed graphics to explain stats -> options to download graphs or images
- Graphical representation + tracking of #hashtags
- Layout manipulate
- Avatar better than user names
- Mode switching: – hide retweet and/or favoured
- Tool to support use of Twitter at conferences – Maps – theme analysis – graphical representation
- Option for infographics rather than just numbers/graphs
- Something like a calendar view with colour coding for # of scheduled tweets/types of tweets i.e. informative vs. fun (etc. more tweets more dark)
- “At a glance” view of schedule of tweets to quickly see how many are scheduled per day
- University/Research Institute should have a central hashtag for announcing events and feeding into topical conversations
- Track engagement with hashtags (events + publications) -> who engages, when, how?
- Useful apps: “it this then that”, ifttt; Hootsuite; Tweetdeck; auto retweet of another account; storify, lists, tweetdeck, tinyurl, pinned tweet; Twitter’s own analytics tool (data not easy to export)
- Identity is important – lots of ‘cheeky’ use
- Way to manage negative aspect/experiences
- Communicate success stories so that others can learn from them
- Manage outreach between multiple research interests + audiences
- Mix informative tweets and fun tweets to inform but also entertain
- Twitter has no way to split work and private lives, so be mindful of the fact that people who follow private twitter also use it for work purposes
- Most current sentiment analysis is not reliable
- University/Institute/Group social media accounts are often run by multiple users who manage a single account. It would be good to be able to (invisible to audience) tag tweets with team member name. Then produce stats on who did what. Also useful for replies i.e. who wrote original tweet?
Please feel free to use the comments option below to add your own thought and suggestions concerning the type of features you would like to see to help you get the most out of your social media use.