Friday, July 6th, 2007
A few weeks ago I wrote a post over on Conversation Hub about participation inequality and the “1% Rule”, which states that for any community, one per cent of your users will participate fully, nine per cent will participate infrequently, and 90 per cent will lurk. My suspicion is that the numbers are variable, and that you would end up with a higher percentage of people participating within a private or semi-private group, particularly those within business. I currently have no numbers to back that up, but I’m going to look for some.
But despite the actual figures, I think participation inequality is not just inevitable, in some cases it’s actually desirable.
First, the inevitability. Communities have always suffered from participation inequality. Long before the internet came along, we saw participation inequality all over the place. Not everyone in a geographical community, or community of interest, could or would take part in the running of that community - people have been disenfranchised for reasons of gender, class, education, religion, affiliation (or lack of), and pretty much any other reason you can think of. Whether it’s the aristocracy, the Old Boys Network, OxBridge or any other sort of ruling elite, we’ve had inequality in offline communities since the dawn of time.
And this isn’t just about politics, but also about science, the arts, even sports. Every field of human endeavour has suffered some sort of participation inequality, and the definitions of who could take part had little to do with ability.
Online, of course, the playing field is much flatter. Most people just don’t care if you’re an Oxbridge graduate or if you left comprehensive school at 16. Some people - like Andrew Keen - get hung up on models of authority, but generally if your point is valid, it’s valid. Online participation inequality is a lot less about tertiary attributes than offline, although there are still plenty of examples where gender, skin colour and sexuality sadly still play a part in some people’s definition of who is qualified to do what. (The online environment is, after all, a reflection of offline society and we still have prejudices that we’d be better off without.)
The inequality that Jakob Nielsen writes about is, I think, a different one - it’s more about choice than exclusion. Ninety nine per cent of people choose not to participate heavily in social activities available to them and ninety per cent choose not to participate at all, and it’s a legitimate choice for them to make.
It would take some investigation to find out why people make that choice, whether it is disinterest, feelings of exclusion, lack of time, etc., but the deliberate exclusionary tactics used by people offline to bar people from a community just don’t work online. I have plenty of online friends whose skin colour, religion or sexual preferences I have no idea about, and nor should I - it’s entirely irrelevant to the conversations we have. Generally I know people’s gender, but not always.
So, if people are, on the whole, making a choice for themselves not to participate, is participation inequality actually a problem?
The idea goes, for businesses, that if people are participating in a branded website, they become more emotionally (and sometimes, intellectually) involved and therefore more likely to buy. Participation increases traffic, and provides value back to the customer. It also gives people something to talk about, which then provides you with that holy of holies, word of mouth promotion.
For many businesses, a social aspect is a good thing to have, but their businesses run fine without it. And for those who do have some way for people to participate, they don’t actually need everyone to do it for it to be helpful. Only a fraction of Amazon’s customers write reviews, yet Amazon thrives. Indeed, Amazon thrived before it started doing reviews because it gives people something that they want.
But not all participation is created equal. The quality of participation is far more important than the quantity. Sites like YouTube attract some very poor comments which do nothing to create or cement a community, or inform or entertain the readers. Most of it is, sadly, dross and this is a common problem across high-volume, low-social cohesion participative sites. Indeed, some communities get positively poisonous. Having lots and lots of comments is not a sign of success if those comments are racist, sexist, homophobic, ad hominem, or just generally obnoxious. It doesn’t help your brand, and it doesn’t encourage the ninety percent of lurkers to either participate, or look well upon you.
Furthermore, could sites cope if participation ran at at one hundred per cent? If you are getting 100,000 unique users a month, and each person left, say, 10 comments, could your system really cope with processing a million cgi scripts? That’s 22 cgi scripts a minute. That’s a lot, especially as, for many of the blogging systems used for participative sites, the choke point is cumulative - you’d be ok for a while, then the whole thing would fall over.
(Note: Of course, visitor numbers follow a power law distribution - many sites have low figures, some sites have very high figures, so I picked 100,000 because it’s a nice round figure not because it represents anything.)
Some sites are set up to deal with volume - Flickr deals with about a million photo uploads a day, but it’s designed to. That’s what it’s there for, and they work hard to make sure that they can cope with demand. Most businesses don’t have the infrastructure to deal with all their visitors participating, whether it’s leaving a comment, or uploading photos, video or audio. The tools just aren’t designed for it.
Beyond the sheer volume of contributions causing technical strain, there’s also the issue of moderation. Libel laws in the UK are very strict, and many online community managers, particularly for commercially-run sites, choose to moderate every comment or item of content. This is expensive, even when you’re only looking at one per cent engagement. Full participation would make the moderation of content functionally impossible and economically impracticable.
And finally, the issues of the breakdown of the community. We’ve all heard of Dunbar’s number, the “theoretical maximum number of individuals with whom [a person] can maintain a social relationship”. Thought to be 150, it has profound implications in just about every walk of life, but it’s especially relevant to online communities where the social ties between participants can be very weak indeed, and prone to breakage. “Communities” of 100,000 people are just not viable - they need to be broken up into much smaller groups in order to really be communities, rather than just a big melee of random strangers.
Inevitably there will be a sweetspot, where you have enough participation to keep the site vibrant, and not so much that the whole thing degenerates into a slanging match. Where that sweetspot is will depends upon the site, the people running the community, the people in the community, the technology, and a host of other things. For some sites, one per cent might be it, for others, ten… or 0.1. I don’t know that there’s any way to predict it.
Yet, there are times when it is incredibly important to be aware of participation inequality, and to actively seek to remove it. When a business website, such as Amazon or YouTube, has only one per cent of its users participating, it’s not a big deal. It doesn’t matter that only a minority of people want to write reviews of books or leave comments on videos. But sometimes it really does matter if only a minority takes part.
The day after I posted on Conversation Hub, Steve Peterson wrote about the same problem on The Bivings Report, citing an example where participation inequality had a detrimental effect:
A recent example of participation inequality side effects is when a Utah school voucher bill was debated on the legislative wiki Politicopia. Utah State Representative Steve Urquhart — and voucher bill sponsor – launched the wiki earlier this year. With great fanfare from publications like the Wall Street Journal’s sister site Opinion Journal, many observers were excited to see how the debate unfolded around the school voucher bill; it faced an uphill battle since similar bills failed during the last several years.
In fact, activity on school vouch[er] bill page on Politicopia is what many consider the reason for its passage. Citizens upset that the school voucher bill succeeded — diverting state money from some of the lowest funded schools in the country — successfully collected enough signatures to have a referendum during a special statewide election in November to potentially overrule the legislature and reject the bill.
Although a group of Utah citizens did participate in the school voucher bill debate on Politicopia, the zeal and excitement surrounding the community was misinformed since participants were a small minority of Utahns. They simply did not accurately represent their fellow citizens.
This problem is one that cannot be ignored. When policy is being created, it is absolutely essential to make sure that it is not based on the opinions of a minority, as happened in Utah. The answer is not necessarily to ensure that every stakeholder should get a say, as this can - managed badly - result in a complete mess. Instead, policy should be based on evidence, and every stakeholder should be able to give evidence.
It’s important to note, however, that the referendum voters in Utah may not have made the best decision, even after the referendum. A process which starts off with one vocal minority and is then changed by a campaign and a popular vote is not an evidence-based process, and there’s plenty of opportunity for bad legislation to be achieved in this manner.
Simply slapping up a wiki and basing your policy on the opinions expressed there is foolhardy and irresponsible. Public debate, e-democracy and the empowerment of the electorate are essential in a modern democracy, but policy cannot be made just by those who have the loudest mouths.
What neither Jimmy Wales nor anyone else owns up to is the sheer exhausting corrosiveness of having to fight with obvious psychopaths like “Quatloo”. I’ve joked that Wikipedia’s tagline ought to be “The online encyclopedia that anyone can edit, so long as they’re willing to devote hundreds of hours of energy to fighting people with autistically long attention spans.” Until Wikipedia can figure out some way of reigning in the rule of its Red Guards, it’s going to be repellent to an enormous swathe of humanity: the people who are put off by authoritarian pricks.
If you want the genuine output of the whole world’s input, you need to stop empowering the volunteer hall monitors over every other kind of human.
Using social software to understand the needs, views and opinions of the community is a valuable tactic, but it must be a part of a wider evidence-gathering process and efforts must be made to involve those who might otherwise stay silent. The dominance of the loudest voices absolutely must not be allowed - they distort the discussion and imply that consensus is all you need. In fact, consensus is not the aim. Good policy based on evidence is the aim.
But when it comes to business, I’ve yet to see a compelling reason why it is inherently a bad thing for only a few per cent of your visitors to engage socially on your site. It is, ultimately, a balancing act between keeping things vibrant and keeping them civil and manageable, and at the moment I think we lack the data to say where the sweetspot is.
I, for one, am thankful that there’s some asymmetry. I don’t think I could cope if every reader of Strange Attractor decided suddenly to strike up a conversation! (And we’re only a little blog!)