Helping ourselves, helping you

Shaun Inman created Mint years before we had GetSatisfaction. He likely got tired of answering the same questions over and over again, and added a bit of instruction to his home-grown forum (below).

Forums don’t scale well

Frankly, if you use a plain-vanilla forum for customer support, you’re sabotaging yourself. When a forum has thousands of threads and tens of thousands of posts, no reputation infrastructure, no tagging, poor input categorization (is it a question, a suggestion, a problem – just what is this piece of text?) and poor sociality, it collapses.

If you’ve noticed how most forums have a) “sticky” threads that stay at the top of the thread list and b) repeated messages admonishing people to search first, ask questions later, well, then you’ve seen why forums scale so badly.

Let the network create value

If we ask make people format their input, it’s easier to manage.

Questions need answers, problems need solving, ideas need discussion and praise deservers response. This one single design decision divides your community’s input into input types you can prioritize.

To limit content duplication, a search is performed and in many cases, your customers get the answers they’re looking for in a step or two. Why? Because often, we have the same teeny-tiny, highly-specific problems. Once the answer has been found, it’s just a matter of efficiently distributing it.

A while back, I asked a support ninja at OmniGroup why they use “traditional” forums instead of GetSatisfaction. I expected them to say “the switching costs are too high” or “we prefer to control our own content”, but their reply was far more interesting:

“I can tell you that since we already have an existing support structure, these services become additional sources we have to monitor, duplication a certain amount of effort. I can see them being useful for a small developer who doesn’t have an existing support infrastructure, though.”

I can see his point: simply adding yet another channel for support would mean more noise and more work, not less.

The Golden Age of Forums is at an end

Using a plain-vanilla forum was once the best we had, but forums rely on our attention, participation and constant gardening to remain vital and leverage the knowledge that’s locked inside the participants’ conversations. That’s because they think in terms of audience and administrator. Some forums do better: Slashdot automatically identifies reputable users and lets them moderate the moderators, as Yochai Benkler documents in the Wealth of Networks. This “crowdsourcing” of moderation work is worth noting.

What makes Community tools productive?

So, how can we transform all this attention, discussion and knowledge into a peer-driven, pull-based knowledge system? I think the points below are vital to customer support communities, but – as we’ll discuss at a later date – they are how we tilt any type of social software away from chatter and towards productive output.

Offer a clear frame for collaboration

• Making people describe their text : “Is this a question, an idea, a problem or praise?”
• Giving topics descriptive names
• Offering a description, with images if necessary
• Tagging topics, so that they can be found
• Describing one’s mood (frustrated, angry, elated, etc)
• Stating that “I have this question/problem/suggestion too”
• Marking items as answered (this does not bar participation, however)
• Showing that this is the main place where the provider and the customer interact helps funnel queries into this one, highly-scalable channel
• (None of this works unless you’re willing to be present and participate in your own community)

Leveraging the cognitive diversity of the community/user base

As Scott Page points out in The Difference, there are four areas where we can make use of cognitive diversity:

• Diverse Perspectives: ways of representing situations and problems
• Diverse Interpretations: ways of categorizing or partitioning perspectives
• Diverse Heuristics: ways of generating solutions to problems
• Diverse Predictive Models: ways of inferring cause and effect

Now, if we assume that the tool has the right kind of sociality, that users are (and are gently reminded to be) civil and benefit from a good standing among their peers, we can make use of their perspectives, interpretations, heuristics and predictive models to discover, understand and fix problems. Of course, built into this is information aggregation and disintermediation. People converge around a topic, and off we go.

Lowering coordination costs

• The above is dependent on very, very low costs of coordination. Imagine doing this by post, e-mail or, ahem, traditional forums.
• When coordination costs are near-zero, everyone can gather on the same page, share input, suggest solutions, share experiences and find answers together. That’s why it’s so important to prevent content duplication – the conversations would be split, limiting the group’s potential effectiveness.

Making topics dead-simple to find

• None of this works, of course, unless people can come across the discussions they need to find, quickly and easily.
• StackOverflow, GetSatisfaction, UserVoice and others assume that people go looking for answers, not their supplier’s support forum, so their threads need to be exceptionally easy to find.
• That’s why, of course, we ask the people who have questions, ideas, problems and praise to tell us what their thread is about in plain language, so that their peers can find them with ease.

Questions? Ideas? Suggestions? Tell me in the comments below. I doubt our conversation will grow too long, but if it does, I’ll embed a GetSatisfaction forum below   And: if you disagree with me and think that forums have a bright future, please tell me why!