Acronym

Like many things social media, this very hyped announcement by the Library of Congress strikes me as, well, not that big a deal. The feeling I got when I read the announcement (other than "does this mean I can get my #bhsm tweets now?") was I guess the Library of Congress wanted some media attention or wanted an image makeover or something. The whole thing strikes me as mostly a PR thing for the Library of Congress--sort of the equivalent of a PR campaign like "See--we're not just about books" or something extolling the modern-day relevance of what some may consider to be an archaic institution.

If they plan to provide some way to extract search results that you can't get from Twitter's own search, then I'd say the archive has tremendous value. If it's just a graveyard of every tweet ever tweeted, I don't really see the value.

Honestly, Twitter's real-time search is great, but as far as not going back further than a few weeks or months, it's kind of useless. I think the real "value" (for lack of a better word) could come from charting things such as trending topics - especially now that Twitter has started local trending topics. It's kind of a barometer of what's important or popular, and could perhaps be interesting in a sociological way.

Some of the attempts to create knowledge out of collective tweets are subject to the same issues we've dealt with in research (and governance) over the past 50ish years. In many areas of decision support we treat our board and committee members as proxies for being able to engage the membership and customer base at large, and when we do put the time into formal research, so often the reader will leap past the aggregate responses (which, sure, can be dry) to read the line by line comments of respondents. The latter can be more compelling because they have a better narrative thread and they are are written by the real customer we're trying to understand, but it's also a misleading practice. It was interesting to see some of the examples of tag clouds offered at Digital Now from the Google & "Supercruncher" speakers and to see, yes, there are some words that are really big and it's cool that this can be generated instantaneously, but it's still no substitute for a thoughtful analysis of the comments, quantitative findings, or back to our point, a twitter stream of 2700 lines.

I also don't mean to paint these as equivalents: the tweets will be valuable for mining individual, interesting observations, but shouldn't be taken as the indicator for mean tendencies. If the room with 100 buns in seats and 25 distinct conversations covered the spectrum of attitudes and perceptions, then a careful analysis of their utterances would be uniquely valuable, as we're getting into the deeper motivations and a real-time stream of reactions and thoughts. But unfortunately it's a self-selected sample and remains a relatively small subset of the overall audience, and it's not necessarily a case of "quality, not quantity" either ... we shouldn't confuse early or timely adoption of new technology for thought leadership per se.

Sorry for the Ben Stein-like overanalysis, but felt it was worth noting now and again as we try to make the best use of what's going on between our peers' ears!

Thanks for the comments everyone. Good insights.

@Maggie: You mention being able to get your hashtagged tweets from your conferences, etc. I'm curious what the specific value is that you're getting out having those tweets collected? What exactly are you doing with them?

@Shannon: Here's an idea for charting trending topics. If you could track topics that are important to your association, and let's say you notice that there's a spike in Twitter activity around a certain topic every Tuesday, or every June 12th, or something like that, you could aim to Tweet from your association on that topic heavily during that period to try to capitalize on the heightened attention. No idea how useful that might be or how difficult it might be to analyze the data to reach those kind of conclusions, but it's a thought that popped into my head just now.

@Kevin: You make an excellent point about whether Twitter is representative of a full population (or membership, in an association's case). For example, at the Great Ideas conference, we tracked tweets and saw that 186 people used the hashtag (about 30 of whom were not at the conference), out of just over 500 in attendance. So that's a sample size of 30 percent or 40 percent of the event population (depending on how you want to factor the nonattendees using the hashtag). That's a pretty good portion for any kind of research, I'd guess, but the self-selection problem is still there.

It's hard to ignore that volume of information now that it's available (whether we're talking about the LOC global level or the small-group level). We're all staring at this information and scratching our heads, thinking "There has to be something we can do with this, right?" I guess for now we take the cryogenic-freezing approach: Capture all this info and preserve it, with the hope that someday in the future we'll figure out how to bring it to life.

There's an interesting article in Slate today that talks about the value historians think they'll be able to get from the archived Tweets. They argue that there's more value to those "what I had for lunch" tweets than you might expect: "'Historians are interested in ordinary life ... And Twitter is an incredible resource for ordinary life.'"