Clever Community Managers + Smart Algorithms
Online Communities are changing. They’re getting larger, sprawling and multifaceted. They’re also getting smaller, and more specific. They’re become more diverse, interconnected, multiplatform – and the tools needed to interact with, find and manage these communities are having to change with them.
But hey, Community Managers are busy people. Where’s a busy CM to focus their attention? Hey. Let me help. How about…
Network Analysis
Social networks are networks. Flattening them with proxy measures like followers, comments, retweets and the like doesn’t actually tell us much about the interconnected web of relationships that governs how communication travels among your community. Network visualisation can show a community’s shape, find the “hubs” that help to propel a message towards virality, identify the “connectors” that form an integral link between two or more sub-communities, and help us identify the influencers which will help a community manager get a message across.
In short, by treating social media as a network, we can find some really interesting stuff. Need more convincing? that illustrates our approach. You could also check out our map of London’s Top 1000 Social Media Influencers.
Machine Learning
We have to deal with a lot of data in social media. Not only that, but we also have to look for quite subtle things: complaints about a brand or the intent to purchase a product, for example. Or how to respond to a crisis situation. Size and complexity typically work against each other: computers are good at coping with volume, humans are good at discerning and acting upon subtlety. Neither is equally good at both.
Machine learning is the best compromise currently available. It is near-human intelligence, at scale. In a nutshell, machine learning is the training of artificial intelligence on a large dataset such that the software gradually “learns” to recognise patterns and is consequently able to make predictions on that data. In an even tinier nutshell: it’s how Amazon and Netflix’s product recommendations work. Ever wonder why they were so damn accurate? Now you know.
But what does this have to do with community management and insights? Well, instead of predictions, the algorithm can be trained to recognise useful patterns in the flood of social media conversations happening in a community or interest area. And these aren’t simple keyword searches. These are complex, ephemeral concepts that are difficult to nail down with traditional keyword monitoring. Imagine being able to track in real time the question “Who is comparing my product with another and expressing purchase intent?” or “Who’s having a negative customer service or technical issue right now?”. Then imagine the benefits of engaging with those customers.
In (yet another) nutshell: machine learning empowers a Community Manager to channel the deluge of public social conversation and engage where they can have the most effect. And that’s a big deal.
Text Mining
The day to day aspects of being a good CM are often concerned with looking after the needs of the community being served. And that’s as it should be: reponding sensitively to the whims of individuals is key to ensuring the group as a whole feels they are being listened to and supported. However, in the context of responding to individual conversations and events, sight of the underling trends within the community may be lost. What topics consistently emerge in discussions? Do particular content types outperform others? Are there sub-communities emerging that require distinct content plans and engagement approaches?
Text mining (also called natural language processing) can help uncover those insights. By looking for patterns or outliers in the occurrence of words and phrases, it’s possible to find meaning in the seemingly random clutter of conversation that can occur on a Facebook page or as @mentions on a twitter profile. Perhaps there are recurring consumer questions that are being lost in the hundreds of comments being left every hour. Or perhaps there are interesting discussions happening about your brand that you might have missed – and these could roll perfectly into your next content plan.
In summary
These are just three approaches to insights and community management. Do you already use any of the above in your role as a CM or Insights Analyst? Or perhaps you would like to, but aren’t sure where to start? Feel free to . Let’s talk nerdy.
Mick is Tempero’s resident Senior Insight Analyst and computer science nerd. In his past life he was a former journalist for Wired Magazine and GQ, and as a surfer hailing from Australia, has survived not one, but two, shark attacks.
Tempero on Google+
