Gathering Evidence on Personal Use of Social Web Services

A recent blog post on “My information consumption habits or how having a smartphone changed the way I work” by Aaron Tay alerted me to tools which can be used to provide an insight into personal usage of various collaborative and communications tools. As the title of Aaron’s post suggests such analyses can help to confirm, or perhaps identify, changes in personal working practices.

Evidence of My Use of Twitter

We may think that we know how we use various tools, but might we be mis-remembering? The MyFirstTweet service tells me that I posted my first tweet on 14 March 2007 : a boring post about filling in my expenses – just like everyone else, I had no idea of what Twitter was about and what benefits it might provide. However I might like to think that I quickly spotted Twitter’s potential and have been a regular user since then. However the Tweetstats service gives a different picture:

Interpreting the Evidence

It seems I made little use of Twitter during 2007 (peaking at 25 tweets in August 2007). It was only in January and February 2008 that I made significant use of Twitter, with 105 and 130 tweets. But this was not sustainable and there were no tweets in the following three months (although this, I subsequently discovered, was incorrect due, I suspect, to a failure for the tweets to have been archived).

Ignoring the uncertainties of my Twitter usage over the missing period it seems that regular Twitter postings began in July 2008 – and from the archive of my tweets on the Backupmytweets service I discovered that this seems to have been when use of Twitter at events and event hashtags was starting to take off in the JISC environment: “AT the JISC Innovation Forum, Keele Univ., listening to Sarah Porter. #jif08“.  And looking at the Twitter statistics for my colleague Paul Walk I see a similar trend, with little usage in 2007, but growth beginning in February 2008, around the time that I started to make significant use of the service.

Incidentally there was a gap in the data for September – November 2008 which made me suspect that my apparent lack of usage from April – June 2008 was due to a glitch in the system, and this was confirming by looking at my Twitter archives from which I can see that I had posted to Twitter during these months. Indeed April 2008 was the month I attended the Museums and the Web 2008 conference and first started to make intensive use of twitter at and event, as illustrated by my social tweet after arriving at the conference. So having started writing this post based on an assumption of the importance of gathering evidence I’m now having to flag the fact that evidence can be flawed (I assume the missing data might be due to the teething problems Twitter servers experienced due to growth in usage).

Since 2008 I have tweeted every month. But this evidence suggests that for over a year after first using Twitter I hadn’t found a particular use for the service. Perhaps this is likely to be the case for other social networking services – there is a need for there to be a significant user community before the benefits can be appreciated. Or, alternatively, perhaps there was a need for better Twitter tools to be developed. Initially I made use of the Web interface, but in July 2008 I was mainly using the Twhirl desktop client and by November 2008 TweetDeck was my preferred desktop client (and, from the archive of my tweets I found that on 8 July 2008 I commented thatTweetdeck 0.15.1 beta is much better than 0.15 🙂“).

Reflections on Implications For Use of Other Social Web Services

The above graph suggests that in the case of Twitter it was only after two years of first using the service that it became embedded in my working practices. I wonder if this pattern will be reflected in my uses of other Social Web services. And if this pattern is replicated across other early adopters of services what might the implications be for the service providers? Perhaps such patterns will demonstrate the importance of building a critical mass of users quickly and the need to ensure that funding from venture capitalists is available to fund the service while its usage if still low?  But what of developments funded in the public sector? Is a two year funding cycle which may be typical long enough to build up sufficient momentum to demonstrate the value of services whose effectiveness may be dependent on large numbers of users?