A Systematic Destruction Of The 90-9-1 Rule

At some point, 90-9-1 crossed over from a Wikipedia observation to a rule defining participation in any community. It’s a lie.

For community building, 90-9-1 is irrelevant. At best, it showcases poor practice or lack of understanding. At worst, it lets you feel happy about 90% of your members not participating.

Here are 8 good reasons why this rule is flawed.

1. Nobody can explain it. There is no explanation. Why would only 1% of people contribute? What is the psychological basis for just 1% of people participating? Does anyone decide not to participate in a community because it’s reached it’s 1% quota? Of course not.
2. Size does matter. 90-9-1 doesn’t apply for smaller communities. Communities of 8 people have near 100% participation rates. Even with larger communities this is a lie. Facebook, World of Warcraft, XKCD and many communities with hundreds of thousands of members surpass the 1% contribution rate. This rule gets broken a lot.
3. 0.001 to 20%. The rule varies more wildly than one might expect. Everything from 0.001% contributions to the Pareto Principle of 20%. That’s a huge unexplained variation. That’s a huge difference.
4. Content channels are not a community. Wikipedia is not a community, nor is YouTube. They are content channels. The difference is huge. You don’t join the CNN community by watching CNN, nor your local community by reading the newspaper.
5. A community isn’t it’s total registered members. Someone driving through the neighbourhood isn’t a member of the community. A community is solely the number of active members. You can’t expect everyone to be a member of a community for life, people come, go and leave. The only difference online is they don’t delete their accounts.
6. Most community managers are awful. Sad story, but the same people that cite the theory to clients are usually those who do an awful job. They treat the community as a homogeneous mass and not the sum of it’s sub-groups. You need sub-groups for a community. You need to cultivate sub-groups between a size between 6 and 15 members for maximum participation. There is research to support this.
7. A community is active people, not passive outsiders. How many of your friends aren’t active in your social group? Probably none – or they wouldn’t be your social groups anymore. They would be people you used to know. A community isn’t the number of registered members, it’s the number of active members. By definition alone, 90-9-1 can’t exist. You can’t have a community without active members, the community is the active members.
8. 100-0-0. This is my observation for the Chess players in Bikinis community. 100% lurk, 0% edit and 0% participate. Do you see the problem? There isn’t a Chess players in Bikinis community. Because no-one contributes there is no community. If no-one contributes you don’t have 100-0-0, you have nothing. Therefore all lurkers, all the 90%, are irrelevant.

Participation inequality does exist in communities. Some people contribute more than others. We need a rule that defines the levels of activity within those that participate, not between those that do and don’t participate.

Sadly, I don’t think it will be as catchy as 90-9-1.

Comments

[Josh Gordon]

Great insight. It’s always good to reexamine assumptions.
Hey now that I contributed am I Now part of the 1%?...Just kidding ;)

[Tom Ewing]

Good post. Your basic point is right I think - something that started as a catchier name than "participation inequality" and less technical "power law distribution" turns into a 'rule' and people measure themselves against it, which is silly.

I'd also add that the 90-9-1 rule gives the impression communities are static - people treat them like segments, whereas the ease and amount of flow between the levels of community is an important metric in its own right.

I don't really agree there's no psychological explanation for participation inequality - it results from a range of behaviours all of which have plausible psychological roots.

I think the danger in the "90%" (inactive members, lurkers, call them what you will) isn't that these people shouldn't 'count' but that it's a catch-all for a huge range of behaviours: from signed-up once and forgotten, to lapsed, to driven off, to reading-but-not-writing, to sharing-but-not-writing, etc etc. There's absolutely no benefit in lumping all those together, whatever your aims are.

[Richard Millington]

I think you're referring to the bystander effect. The more people present, the less people are likely to get involved.

But that really doesn't apply here. You don't know how many are participating. In any community, it's hard to tell if it's 1% or 50%. Unlike the bystander effect, you don't see the other bystanders.

If anything there should be MORE participants because you only see people that are participating. Those are the only people you can copy.

Since it's impossible for anyone to work out the % of people participating compared to the 'total audience', the % cannot impact the number of people who participate.

[Liam Barrington-Bush]

I like the point about the 90-9-1 rule seemingly 'letting people off the hook' to describe a niche (online) gathering amongst a few friends as 'a community'. Just because someone signed-up once is no reflection that they are a 'part' of your community, any more than someone who visited a foreign country once automatically becoming a national of their vacation destination...

It will be interesting for me to see how some of these ideas play out on the campaigner website I've been developing, http://www.Louder.org.uk - as there will be an online community component to the site, but will be primarily about providing individual members with the online tools to take independent action for their causes... but then to be able to connect and learn from each other as the site develops... Thoughts are always welcome!

Good post! Interested to hear more on the topic!

[Tom Ewing]

I wasn't thinking of the bystander effect, tho on the net the sense of an invisible/potential audience can cut both ways I guess: it's the motor that encourages people to do things, so it wouldn't surprise me if it also intimidated people. Site users don't KNOW the 'total audience' but they have an approximate idea, and that's bound to influence how they relate to and interact with something.

Anyway, I was more thinking of territoriality and deference as the motor: people who feel the community is their territory more likely to post than people who feel it isn't. So your point on subgroups is a good one.

Also, does there need to be a psychological explanation? ;)

[Richard Millington]

I disagree 100% with the 1st paragraph and agree 100% with the 2nd there Tom.

I can't imagine how the number of lurkers would impact whether anyone decided to participate in a community.

Using Wikipedia as an example. You don't decide not to edit/contribute to Wikipedia articles based upon the number of contributors to lurkers. You have no idea of either. You contribute if you have a good reason to. You contribute if you care about the topic, or you care about being in the community of Wikipedians.

I think a psychological reason is needed. Rules have explanations. People do things when their brains tell them to. If you decide to participate or not, there is a reason for it.

[CoCreatr]

Thank you for the round-up of a lot of sense. I see a big difference between communities people join because they want to and plain sign-ups (or a meaningless number of remote "friends" on social networks). Does this call for a clarification of the 1% Rule for virtual communities on Wikipedia?

http://en.wikipedia.org/wiki/1%25_rule_(Internet_culture)

[pfctdayelise]

Er, how is Wikipedia not a community?

No, it's not a community for everyone that uses it, particularly if they only read. But for editors - why do you say there not an editor community?

[Richard Millington]

You're right. Wikipedians are a community. They are a community of people that do edit and contribute Wikipedia articles.

I didn't mention them because I didn't want to confuse the issue. People don't refer to Wikipedians when discussing the 90-9-1 rule. They refer to every single visitor to Wikipedia.

Which is madness.

[Joe Cothrel]

The "1% rule" and 90-9-1 are two of the most useful rules for community practitioners, so I feel obligated to chime in on this less-than-systematic mess you've created. Let me see if I can do this in as few words as possible.

I see here a few things that are right, a few that are wrong, much that is confusing, and one or two things that are wrong via exaggeration.

#1 and #3 are just errors in logic. That fact that you can't explain something doesn't mean it doesn't exist. The fact that there are exceptions doesn't nullify a rule. Basic stuff.

#2 and #5 are largely correct. Credit for that.

#4 is just confusing, because you never define your terms.

By the way, it would also be nice to know what size and type of communities you base your conclusions on. Some of what you say strikes me as right for smaller communities but wrong for larger ones. To put my own comments in context, my experience is with large communities (>5K monthly users) created by companies for their customers -- about 300 communities in total over 10 years.

#6 may be true to your experience, but it's certainly not to mine. Although, if most of the community managers you know are "awful," perhaps that explains some of your ideas.

As an aside, your sub-groups principle has never made sense to me. Based on communities I've seen use that technique, it's just as easy to kill a community that way as it is to grow one.

#7 is also confusing. You seem to be unaware of the difference between a community and a group.

#8 seems obvious.

Your first two paragraphs consist entirely of statements that are either a) wrong or b) partially true but exaggerated to falsehood. 90-9-1 didn't start as a "Wikipedia observation," it started here, with Jakob Nielson:

http://www.useit.com/alertbox/participation_inequality.html

(Many people make the same mistake. I think you just fell prey to Jeff Howe's rather relaxed attitude toward attribution.)

More important, while you cite some exceptions and qualifications, you never demonstrate that 90-9-1 isn't generally true -- but you still insist it's "a lie," "irrelevant," and an excuse for complacency.

Exaggeration is, of course, very common with blogs striving for popularity. But it's less common among bloggers whose goal is to share knowledge, demonstrate expertise, and advance an area of practice. That's because when you exaggerate, you lose credibility.

[Richard Millington]

Joe,

Thanks for the comment. Lets leave the semantic issues out of this. Defining 'community', 'groups' and 'content channels' would send us in circles.

I agree with some of your points. Exceptions don't nullify a rule, nor does a lack of explanation. It should certainly bring the rule into question though. Every exception questions the rule. The lack of explanation should challenge the circumstances that provoke the rule.

What are the boundaries of the rule? Does it define community of 1000? 5000? 500,000? What participation % will you get from a community of 500,000 compared to 1,000? What are the links? Why does the size affect the number of people who participate?

These are key questions the rule doesn't answer.

Sub-groups are a social observation, by which it's impossible for 5000 people to feel part of the same group. A person can't have 4999 friends. Instead sub-groups (friendship circles) form within that 5000 which allows everyone to be involved.

I can't prove 90-9-1 isn't true because it is. It's true in certain circumstances, if you take certain actions, if you have a different understanding of a community (as we probably do) and if you run a community in a certain way.

But that doesn't make it a rule for online communities, it makes it a rule for those circumstances. This post is to challenge those circumstances, which in turn should challenge the rule.

[Simon Peters]

Maybe it's actually the term community that's causing consternation here. Would the 90-9-1 rule be a lot more accurate if it was worded "10% of an online population will use community tools (in other words, only a minority of a population will form a community)"?

It's a rule of thumb of course (any psychological dictum always is) but the member populations I've worked with have generally followed the spirit if not the statistics of the rule. That is, a minority will create content, a larger minority will amend content and the vast majority will consume content. This generally holds true of an economy too.

The examples you give in point 2 are actually good examples of how to contend with the rule. facebook defeats the rule because the entry process forces you to become a content creator (populating your profile, adding pictures, etc). WOW actually has no content creators or amenders 100% of them are actually consumers of content.

I do get where you're coming from - but the 90-9-1 rule does have legs - it just needs to be used in context of the nature of the population and the relationship between the producers, amenders and consumers in your community.

(Aside: what does everyone put the boot into 'lurkers' (read: consumers), if 1% can produce enough content to satisfy 90%, that's brilliant. I'll take that and call it an achievement any day.)

[Rachel Happe]

Wow - what an interesting conversation. I am in agreement that the 90-9-1 rule is not generally useful... but I think it gets back to the way community is defined which is what seems to be causing a lot of the differences of opinion.

I see it this way:
- Groups are collections of people who have individual relationships with 70-100% of the other people in the group.
- Communities are collections of people who have individual relationships with 30-70% of the other people in the community.
- Networks are collections of people that have relationships with <30% of other people in the network
- An audience is a collection of people who are mostly unknown to each other

Now that is far from a universal definition... it's mine and mine alone. But for me, to have a group or a community with low participation would break its very definition. Also, a caveat which is this. Lurkers or bystanders or whatever you call them, they have value and they do participate. Content has no value without people who consume it.

Where does that get us? A lot of confusion because there are not standard definitions of what it means to have an online audience vs. and an online community. For that, those of us who are professionals in this space need to start helping define and set expectations around these things.