Australian companies have put a great deal of thought into turning customers into free marketing and technical support staff online.
Through Silicon Valley social software vendor Lithium, firms like the Commonwealth Bank and Telstra are analysing how their customers interact online, and ways to turn them into enthusiastic brand advocates.
Lithium chief scientist Michael Wu, who is a biophysicist by training, analyses more than 400 corporate clients’ web statistics to build predictive models of customer behaviour.
He tests his mathematical models against psychological, sociological and economic concepts to identify social media strategies that help Lithium’s clients acquire, engage, enlist and monetise online fans.
Dr Wu shared some of his insights with Business Insider, including the technical principles behind the high-level advice he provides to clients.
Some organisations hire teams of 'social media managers' to respond to customer queries on Twitter, Facebook and discussion boards and drum up enthusiasm for the brand.
Wu says relying solely on social media managers to build online rapport just doesn't scale.
'Mathematically, the number of different conversations between n customers is ½n2. It's quadratic,' Wu says, meaning that conversations grow at a rate that is proportional to the number of customers. A large number of customers means an explosion in the number of conversations.
'But the number of conversations that people can handle is usually linear: it grows with the number of employees. Quadratic will always win in the long run.
'A social media manager in a company can never scale. You can never hire enough employees to respond to all the social conversations out there.
'The only chance that you have is to collaborate with your customers.'
Online communities are described by the '90-9-1 rule': 1% of members are what Wu calls 'super users', who drive about half of all discussions.
Only about 9% of members actively participate in discussions, while 90% of members are 'lurkers' who tend to read, rather than contribute, posts.
Wu says online community participation can be described by a power law function, P(x) ∝ x-α.
He says the 'power law exponent', α, is a measure of 'equality', which reflects the ratio of super users to other users. It typically ranges between 2 and 4.
Wu says inequality drives a more vibrant community because it gives lurkers a level to strive towards.
'Whether or not it's ideal, it's very hard to change. People have tried to change α in some way, but there's some natural dynamic that's built into it,' he says.
'To try to force it either way is not healthy behaviour because you'll have to either cap participation by telling super users that once they reach a certain level, they can't participate anymore, or you force people to participate.
'It creates this tension that is not natural.'
Organisations want super users because they're the ones who are there to answer questions when 'social media managers' can't.
When Telstra released the iPhone 5 last year, members of its 'Crowdsupport' discussion board answered more than 40,000 support questions about the device within a month of its launch.
Wu says there are hundreds of factors that determine whether an individual becomes a super user.
'A lot of people think super users are the ones who exhibit a spike of activity by contributing a lot over a short period of time. That's not true,' Wu says.
'It turns out, the top predictor for super users are the ones whose participation velocity is steady and increasing.'
By that, he means that individuals who consistently contribute to discussions and who become more engaged over time are more likely to become super users than those who join in a flurry of posts.
'In retrospect, it kind of makes sense: you see a huge spike of activity from people when their phones are not working or who want to solve a problem, and once they solve the problem, they go,' Wu says.
'Super users are the ones who answer a question here, a question there - they don't stand out as much.'
Much of Wu's analysis goes on behind the scenes. Lithium clients receive a 'Community Health Index' score that tells them how well their online customer communities are serving the needs of their users.
A high CHI score denotes more engaged users, but not necessarily commercial success.
Lithium's community health score accounts for member registrations (μ), quality of content (U), traffic (vh), interactivity (I), the time it takes for users to respond to messages (R), and liveliness (L).
Interactivity and responsiveness are functions of the number of discussion threads (Θ), the number of users participating in a thread (uΘ), the number of messages in a thread (mΘ), and the time taken for someone to respond to the kth message (ΔtΘk).
The quality of content is a function of the number of new posts (p) and traffic.
Liveliness is a function of the number of publicly accessible discussion boards (B) and the expected number of posts per board (pe).
Lithium encourages its corporate clients to use game-design techniques to keep customers engaged.
That means rewarding online users for certain behaviours, like contributing a certain number of posts to discussion boards - at least until they are self-motivated to continue behaving that way.
'People are initially driven by tangible rewards and later, intangible rewards. That is pretty well known in behavioural economics,' Wu notes.
Based on his analysis of Lithium clients' communities, Wu says companies should structure their membership tiers based on linear increments: if it takes 10 posts to reach level 1, it should take 20 posts to go from level 1 to level 2, 30 posts to reach level 3, and so on.
Balancing difficulty and rewards encourages users to go into the psychological state of flow, Wu says, which is the state in which a person is fully immersed, involved and enjoys a particular activity. Think of it as being 'in the zone'.
'In order for people to go into flow, the challenge they face has to match their skill. If the challenge is to great, they feel frustrated and if it's too easy, they feel bored,' Wu says.
'You want linear increments. The (optimal) equation is essentially quadratic. A lot of community managers who create rank ladders typically create ladders that are either too easy or too difficult.
'It needs to be easy in the beginning but successively get harder. Typically, people go geometric or exponential, but that's too hard - you'd have to do twice as much work to get to the next level.'
The scientific concept of preferential attachment - where the rich get richer, or popular people get more popular - applies to networks like Twitter, but not online discussion boards, Wu says.
That means the controversial strategy of buying Twitter followers or Facebook friends to appear more popular won't help you win kudos on discussion boards.
'Celebrities tend to draw a lot of people who want to connect to them,' he says. 'That works on social networks to a certain extent, but it doesn't work quite well in communities.
'People don't connect to a certain user in a community because they're a celebrity; relevance of content and reciprocity are stronger determinants of who is going to connect to who.
'For example, you may be an expert in iPhones on the Telstra community. People are not going to connect to you because of that. They're going to connect to you if they have a problem and you help them.'
Sociologist Mark Granovetter's 1973 paper, 'Strength Of Weak Ties', underpins many of today's corporate social media strategies.
According to Granovetter, there are four components to a relationship: time, intensity, trust and reciprocity.
'Time' describes the amount of time spent together, not the total length of a relationship, Wu explains.
Intensity refers to the quality of time spent together, trust reflects transparency and openness, and reciprocity reflects how both parties - the company and the customer - have to gain from the customer engaging with the brand's online community.
'When you build these four components individually, you will have a stronger relationship. There are tactics to build these components,' Wu says.
Wu acknowledged people and organisations may appear insincere if they apply social media strategies and tactics carelessly.
'Some people think, oh with gamification you can do behavioural conditioning: I'm going to condition these people to be like pigeons and do these tasks for me. But customers are smart. They're going to know that they're being taken advantage of,' he says.
'With science, we discover knowledge. We codify it. How people or businesses want to use that, it's their choice. We can provide guidance, but ultimately, knowledge is agnostic of how it's being applied.'
Business Insider Emails & Alerts
Site highlights each day to your inbox.