After the discussion with the Road Safety Forum guys I did some chatting with some public service types and turned up a few issues, in particular what can only be termed the ‘lifecycle’ of social media channels.
I did a little google, and couldn’t find anything useful about what I’m posting here, but if anyone knows of some good thinking in this space I’d love to hear it.
What all frequent users of blogs or bulletin boards should have heard of is ‘Godwin’s Law‘. This law states that as a post ages the likelihood of “Nazis” or “Fascism” being invoked increases exponentially (I abridged). The use of these two terms, or a derivative of them, usually indicates the end of useful conversation. Now, blogs and/or forums are all about conversation, so a thread “Godwinning” is effectively a death kneal for a topic as well.
Something I’ve noticed though is that threads tend to reach a natural conclusion of sorts even if Godwin’s Law doesn’t apply. From observing in participating in conversations over at Public Address System I can state that Godwinning very rarely occurs. But, threads do tend to reach a kind of ‘peak’ before useful conversation tapers off into a blaze of many and varied topics and viewpoints.
What this has lead me to think is that a productive thread will follow a particular lifecycle. I say productive thread because plenty of threads are unproductive. I haven’t visited Kiwiblog since moderation kicked in, but back in the day it was highly unusual for a thread to contain productive or useful information. Knowledge generation? Forget it.
And for public servants, social media is partially about knowledge generation. It’s also about fostering communities of interest, social partnership, etc etc, but it’s especially useful for analysts or advisers seeking knowledge.
I know that “knowledge generation” is something of a cliche, but it’s a useful one for what we should be using social media for. This is because there are a hundred ways to gather information. But gathering information in a socially proactive way, and being able to build on it over a given timeframe? Much more valuable, especially to a researcher like myself.
So how does this relate to the lifecycle of a given social media?
All social media suffers from variable signal-to-noise ratios. What I’ve noticed is that early in the lifecycle of a social media the signal to noise ratio is high, meaning that there isn’t much useful information. This could be because there isn’t yet a critical mass of participants, or because the groupthink hasn’t adequately kicked in. But over time the babble lessens and more interesting signals can be gleaned. Eventually though, the coherency and usefulness of the conversation will again decrease.
I should add though, that the usefulness will only decrease to the person who has posed the questions. To the people participating in the conversation the usefulness is subjective, so they could still be enjoying discussing an issue or topic long after the key points have been well and truly expounded.
What’s important to making social media work is that both these needs are considered. The poster or ‘owner‘ of the site needs to be able to get good information or knowledge out of the process, and participants need to feel that a conversation has reached a natural conclusion. The former ensures that the owner thinks they’ve established a good process or tool, and the latter ensures that participants feel like they’ve both contributed and been part of a community of interest.
To sell these ideas to an analyst or adviser you need to ensure that they know that a social media will only have a limited lifecycle, and that it isn’t a long-term or onerous task. Much the same as any period of feedback or information-gathering there is a natural conclusion. And while this conclusion might not be scheduled the way a static or non-social media interaction is, it will eventually close.
The key difference is that social media offers a far richer type of interaction, and an opportunity to include the public in the way direct polling, focus-groups, or interviews cannot.