Subscribe to the Online Community Blog Via Email

Your email:

About Socious Software

Customers Love Online CommunitiesSocious provides online community software and services to companies and associations to increase revenue, improve customer retention, and create more market-driven profitable products. Learn more about Socious online community software.

Posts by category

Online Community Blog

Current Articles | RSS Feed RSS Feed

Is the 90-9-1 Rule for Online Community Engagement Dead? [Data]


There is a rule that has floated around in the social media world for quite some time called the Rule of Participation Inequality or the 90-9-1 Rule. This rule states:

User participation in an online community more or less follows the following 90-9-1 ratios:

  • 90% of users are Lurkers (i.e., read or observe, but don't contribute).
  • 9% of users are Commenters. They edit or rate content but don’t create content of their own.
  • 1% of users create content and are Creators.

This rule gives both hope and discouragement to organizations that are creating online communities. It gives hope to companies and associations who launched an online community and are not seeing any business-level benefits since they believe that 1% engagement is acceptable. It also discourages organizations from exploring how an online member community or customer community can impact their business if they believe that only 1% of their audience will fully engage. If you are creating an online community for people to participate in, and 90-9-1 statistics dictate that a high percentage of people will not participate, where is the benefit?

Having heard this rule for years and seeing what I suspected were higher levels of participation in our customers’ online communities, I began to ask myself if the rule is really true. So, I set out on a quest to see if that 90-9-1 Rule holds water.

Testing the 90-9-1 Rule for Online Community Participation

Is the 90-9-1 Rule Still Valid?

Why Did I Conduct This Research?

If the rule did not hold up, many companies and associations may be damaging their business and marketing strategies by basing decisions and benchmarking results using a general rule created in 2006 . So that the readers of this blog have a point of reference for when this rule for online communities was created, keep in mind that Facebook ended 2006 with only 12 million users (Facebook now has over 650 million users).

Study of Online Community Customers

I compiled statistical data from a random sample of our customers so that I could crunch real numbers to determine if the Rule of Participation Inequality was true for private online communities. To begin, I had to assign actions to measure at each level. So here are the actions I assigned:

  • Lurkers: Have logged in and viewed information.
  • Commenters: Have commented on or edited a blog, wiki or file or have answered a forum post that was already initially asked.
  • Creators: Have initiated a blog, file, wiki or forum post.

The thing about the rule is that it infers that all users are doing something since the 90-9-1 all add up to 100%. The problem is that many organizations have profiles of users that are deactivated, past members, or guests. Also, not all members of an online community have access to the same tools, content, and functionality. So, to make a fair correlation, I ran two sets of numbers - one set accounting for all profiles in the system and one set with only the participating users making up the 100%. These numbers are below:

Analysis of the 90-9-1 Rule for Online Community Participation

Findings of the Online Community Research

My belief is that the second chart is a more accurate comparison to the 90-9-1 rule since all users have to be doing some activity to account for the 100% of the sample. So based on the data in that chart, there are a few interesting things we can learn:

  • All but one online community had more Commenters than the 9% the rule suggests. So, people seem to be more open to editing and commenting on existing information.
  • All sites were higher in Creators than the 1% the rule maintains. One as high as 17%! With more and more people getting comfortable with social networking sites, perhaps people are more comfortable in expressing their opinions.
  • The averages for each area are far higher than the rule suggests. (Well, other than Lurkers, but that is a good thing!)

So, maybe we don’t need to be so dire about how many people engage in your online community. Based on this data I would suggest a new rule (with a little rounding):

The 70-20-10 Rule of Community Participation

Has a nice ring to it doesn’t it!

Online Engagement Takeaway

How would it impact your business or membership organization to increase your online engagement among customers or members who create content in your community by 1000% and more than double the number of people who engage in discussions and comment on content? For most organizations this would mean higher customer retention, better customer satisfaction, and more passionate brand advocates in the market. The data above tell us that you don’t need to settle for generalized guidelines. Look for strategies and technologies that are proven to break long-held rules and benefit your organization.

At Socious, we believe that your customer, employee, and partner communities hold the key to profit and innovation. To help companies get the most from our all-in-one online community software, we have made significant investments in educating businesspeople, guiding customers, and leading the customer community movement.

Please take advantage of these free resources:
Online Community Blog | ProCommunity Interview Series | Whitepapers & eBooks | Videos


Interesting study!
Posted @ Thursday, August 11, 2011 11:56 AM by Giles Rafol
I think this thinking needs to be ratcheted up a few notches. There are, in my experience, 3 different kinds of communities. Communities of interest (more casual), communities of practice (you and I share the same type of "work" and a notion of reciprocity is inferred) and communities of purpose (objective based, not yet another web address, begin and end through a facilitated (architecture and people) process) - all of which have much different participant motives and behaviors. There's a need to have a better informed understanding of the nature of the community and it's spacial architecture so that "studies" are designed to have more value. This 90/9/1 and 70/20/10 is different for the different kind of communities.
Posted @ Monday, August 15, 2011 3:38 PM by Nancy Lewis
Thanks for your comment. I agree that not all online communities have the same membership makeup, goals, and value proposition. Our customers are mainly B2B companies, software user groups, and large associations. By your definitions, the communities I analyzed are both “communities of practice” and “communities of purpose.” These are online customer and member communities designed to help online community member be more successful with their jobs and businesses, while also mobilizing the community to collaborate of product features, legislative initiatives, and solutions to technical problems.  
I appreciate the distinctions you laid out. The spark that started me down this investigation were a series of consultants who gave presentations to business and associations claiming that the 90-9-1 rule was gospel for online customer and member communities. Though I don’t have data to back it up, I have a hunch that “communities of interest” may have engagement levels that are closer to 90-9-1.  
Posted @ Monday, August 15, 2011 5:12 PM by Paul Schnedier
Hi Paul 
I enjoyed your post. 
When we looked at the 90/9/1 "rule" we added a "900" in front, for those in the organisation who never went near a particular space. 
I think the 90/9/1 can be reconciled with the 70/20/10 if if you think of the first as the profile of people engaging with say, one particular wiki focused on one aspect of the business. But if you have many wikis the numbers who initiate articles on one or more increases from the "1" to the "10". 
I also think that these numbers apply to community-wide domains or spaces. Within a community or organisation of say, a 1000 or more people, there will be many spaces open to much smaller groups, where people feel more able to contribute.
Posted @ Tuesday, August 16, 2011 2:56 AM by Nick O'Doherty
Intersting and useful study/informationPaul. 
I'm looking at this from a nonprofit association leaader/executive standpoint. 
I wonder if the engagement using all profilts in the system might represent the entire association, thus for any given assocaition we may be looking at a 33-9-4 rule ... in addition to the 70-20-10 rule for those members who are online? 
Posted @ Thursday, August 18, 2011 8:38 AM by Steve Drake
This is good work, thanks for publishing it's a healthy challenge to the original 2006 principle.
Posted @ Saturday, September 03, 2011 2:29 AM by Michael Ryding
Is someone who hits "share" on or on any of the gazillions of funny picture sites a "creator"? 
Is omeone who answers such a post/article with "funny" a commentor? 
Simply counting numbers in social networks without a qualification of the contents is hardly significant, with thousands of websites having those unspeakable "share" buttons.
Posted @ Tuesday, October 18, 2011 7:59 AM by Markus Stumpf
Lurkers, although not actively participating on the subject online, may however also share a subject with their friends/relatives/collegues during regular offline conversations (adding their own view on it). Those "lurkers" can/will spread and edit the word also. Although this is not directly recognizable, it must be seen as a valuable contribution to get the community growing.
Posted @ Thursday, August 30, 2012 7:50 AM by Frank van Rijt
Thanks Frank. I agree, word of mouth advertising for a product or community is always great. Hopefully the lurkers mention the community they saw the content in to help drive others to it and participate. Thanks for your comment.
Posted @ Thursday, August 30, 2012 10:57 AM by Paul
Post Comment
Website (optional)

Allowed tags: <a> link, <b> bold, <i> italics