Wednesday, May 13, 2009

Managing Twitter accounts for libraries - What are libraries doing?

There is considerable literature by now on handling reference and/or communicating with patrons via phone, email and Instant Messaging, but would the same considerations apply for Twitter?

Still certain issues/questions (if not answers) would be similar for all types of channels.

First off, I must confess, I have not set up a Twitter account for a Library, so this post will have more questions than answers. But I have being looking at what other libraries have done on Library Twitter League to provide some guidance which will provide most of the meat of this post.

Still nothing beats personal experience, so please help by commenting on how you manage your Twitter account for libraries!



Issue 1: What information is going to be twittered? How is Twitter going to be updated? Manually? Automatically? Or a combination of both?

The easiest
way of course would be to leverage on existing work already being done , extract the information into an RSS feed (e.g. From blogs, delicious accounts etc) and feed it into Twitter using tools like Twitterfeed. Alternatively, if one already had a Friendfeed account, one could just aggregate everything into Friendfeed and then set Friendfeed to post to Twitter.

Such fully automated accounts generally have high number of updates and my subjective feeling from looking at such accounts is that they appear to be less engaged with users because the "light is on, but really there is no-one at home"

Such automatic methods alone would also mean a loss of flexibility and reliance on another layer of service (Friendfeed or Twitterfeed) which the library would not control on top of Twitter.


Issue 2: What accounts should one follow? Should one auto-follow every user who follows you? Should one proactively identify accounts belonging to users and follow them?

In theory there is no need at all to follow any other Twitter accounts, as one can freely direct replies denoted by @username to any account even one that one is not following. So for instance if a user directs a query to you (he need not be following you) , you can reply using @username without following him.

There are in fact three reasons I can think of why you might want to follow users

Firstly, once you have followed a user, he can send you a private direct message , and this preserves the privacy of any transactions that occurs.

The second reason why Libraries might want to follow other accounts is to encourage them to follow back! I can imagine a policy that involved doing a search for the institution name, or library to identify users of the library, and follow them in hope of them following you back. Not the best reason of course but still....

The third reason is probably the strongest. If a twitter account follows your Library account, chances are he is a user of your library. Or you might follow twitter accounts if you are sure he is a user whether he follows you or not. As such when he mentions "library" it will automatically appear in your timeline and you will be see it and be reasonably sure he is talking about your library and you have a chance to respond to it.




The example above which shows a threaded reply (via Twitree) that initially puzzled me. While some libraries do environmental scanning and proactively answer queries or complaints when they notice their name or website being mentioned (see examples later) despite not being explicitly addressed, how did "Okstatelibrary" ( Oklahoma State University Libraries) in this case know that the user "hpat" was talking about them?

After all, all that was mentioned was "library" and the query wasn't directed at them.

The answer can be found from the fact that "Okstatelibrary" seems to at the very least routinely follow all twitter accounts that follow them or possibly even follow accounts that they identify as belonging to their users even if they are not following back (implied by the fact that the follower/following ratio is below 1) , so they can be sure that "Hpat" which they happen to follow is a user and is talking about them.

Without doing this, even a environmental scan of your library name would not pick this up.

Does that mean it makes sense to try to proactively identify users of your library on Twitter and follow them?

To see what libraries were doing , I did a rough empirical study.

Based on the top 100 accounts (by follower counts) on the Twitter Library League, I looked at their Twitter Follower/Following ratios, to see what strategy they were following.







As you can see above, 8% of accounts followed no-one at all!

15% of accounts followed more accounts than they had followers. Particularly some accounts had ratios as low as 0.4! Possibly they were trying the strategy mentioned above of following accounts they identified as belonging to their users ? For example the thread below (courtesy of Tweetree) seems to indicate a proactive stance of finding users to follow?




However 29% of accounts had followers that were between 1 to 2 times as many as they were following, and almost half (48%), were followed by more than 2 times as many accounts as they were following.

This seems to indicate that in general libraries were quite selective in following accounts, and most did not auto-follow every account they followed. An open question for further research though is what accounts were followed and why.

This article recommends that you link to resources such as BBC's Twitter account, though a quick look seems to suggest that accounts that were selective in following accounts, generally chose other library accounts, vendors etc to follow!

Further evidence can be seen from the special Twitter account I setup that follows all Library Twitter accounts tracked on the Library Twitter League.

Out of 450 Library Twitter accounts I followed, roughly 68 accounts followed me back (not all accounts following that account are Library accounts as some Librarians probably thought it was interesting to watch me and I didn't feel like blocking them), roughly a follow rate of 15%.

I couldn't tell if any of the 68 was autofollowing me, or had a manual policy of doing so, though a couple of accounts had Follower/following ratios of 1 or less, that seem to indicate so.


Should one send an auto welcome message to all users following you?

Another interesting thing is that I got 5 direct messages from Libraries, welcoming me and thanking me for following them. Probably a canned auto-message? Of the 5, only 2 followed me back though.





Issue 3: Are users sending queries via Twitter? Should libraries proactively try to answer queries concerns them but are not directed at them? What targets should be set in terms of response time when answering queries received via Twitter?

Even if you do not set up your Twitter account to be an out-right reference query service, you will no doubt start receiving queries. I wondered though how many queries one expects to receive. The 140 character limit doesn't make it particularly handy to ask or answer queries ,so probably answers would have to be moved to another communication channel (email probably).

To answer these questions on what libraries were doing again, i turned to the top 100 Library Twitter accounts (by follower count) listed on the Library Twitter list. I used TweeTree (which showed threaded conversations with dates and times) to review roughly 2 week worth of Twits for each account to look at replies sent to and from the Library account.

I unearthed quite a lot of interesting information, so much so that the bulk of it will be the subject of another blog post, I will just mention some relevant details here.

In general how many queries should you expect via Twitter? What type of queries are being sent?

I found that in general the number of queries directed at Library Twitter accounts were quite low. Many accounts had in fact zero real queries (retwits are not counted) by users directed at them. These accounts seemed to be mostly automated accounts, pumping out large number of updates on new book arrivals, blog entries, news etc. Possibly, followers assumed no-one real was behind it, and hence didn't bother to send queries there.

The accounts which were generally more engaging of users got slightly more queries directed to them but even the most active ones didn't seem to get more than 1 or 2 a day. These were typical directional questions ranging from questions about times of events



to queries about contacts




to procedure questions about replacing lost cards



to queries about searching



Should you proactively monitor mentions of your library and respond to them?

“I use Twitter’s search feature to locate people who are in the Grand Rapids area who tweet certain keywords. Words like “library,” “GRPL,” “reading,” “book,” and “Evergreen” which is the name of our new catalog. Once you do the search you can create RSS feeds that are updated whenever someone mentions the words. I then subscribe to the RSS feeds using Google Reader to periodically find new people, insert the library into conversations, and see what people are saying about the library.” -- Source

One thing I noticed was that accounts that got more queries from users were in fact also pretty proactive, doing environmental scans to answer queries and complains that weren't directed at them. Probably such accounts appeared less robotic, were engaged and as such users were encouraged to ask questions.

Some examples.

Proactively responding to a complain because NY Public Library was mentioned.





Proactively answering a twit that mentioned the website name



Proactively answering yet another a tweet that mentions NYPL (new york public library)




What are the expectations of users in terms of response time when sending queries via Twitter?

Given the fact that people tout Twitter as "real-time", would their expectations for replies be closer to that of users communicating via Instant Messaging or would it be closer to that of email (roughly expecting a reply by the end of the business day).


There is no way to tell about expectations without surveying users, but a quick unscientific look at roughly 50 threaded replies from library twitter accounts, seems to indicate that most library twitter accounts were responding to queries from anything between 2 minutes to 4 hours . Almost all were within the day for sure, which seems to indicate that for librarians at least they were treating it similar to queries via email.


Should you arrange for staff to monitor Twitter at the same time as monitoring other communication channels (SMS, Email, Instant Messaging)?

If monitoring Twitter directly via a client or the web is too troublesome, one can set Twitter to send email alerts, whenever a direct message is sent to you. But I believe it doesn't do so for replies to you though, though there are many workarounds to get informed via email, though it's unclear what delay this involves.


Issue 4 : What targets if any should be set?

As detailed in my last blog post, a target of 200-250 followers is probably realistic. If one was ambitious, one could aim for anything between 350-1,000 followers and that would put one within the top 50 Library Twitter accounts (out of some 450 accounts tracked). Anything above 2,000 followers is pretty much impossible, currently only Library of Congress has gone beyond this level, but they are clearly a special case.

Now dear reader it's your turn, if you have a Twitter account for a library, how do you handle the issues mentioned above? What other important issues, points am I missing?



blog comments powered by Disqus

Share this!

Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.
Related Posts Plugin for WordPress, Blogger...