Originally published on September 13, 2016
In a typical community, 10% or fewer of the members will tend to post, ask questions, present, etc. The rule of thumb is that 10% of the members will participate at all, and only 1% will regularly be active in discussions and presentations. 90% will not post or speak up at all. Some have questioned whether this rule of thumb really applies.
I recently received this query: “Do you have any numbers for ideal engagement for internal social networks? Something like this:”
I replied: From my experience, the percentages shown above are not realistic, and the 90–9–1 rule of thumb might actually be optimistic.
This article provides actual data from a social network, a community, and a social media platform. It also quotes and includes links to other articles, including a range of perspectives.
1. Enterprise Social Network Data
Here is the analysis done in a very large enterprise social network (ESN) with more than 130,000 members:
From Lee Romero: “I put together an analysis of public groups in the ESN for January and February of 2016. Some insights:
For January 2016:
- Average lurkers (members who do not post): 96.61%
- Average Active (members who post > 2 times in a month): 0.62%
- Average Mildly Active (members who post one or two times): 2.35%
- Average of members who post anything at all: 2.97%
For February 2016:
- Average lurkers (members who do not post): 96.99%
- Average Active (members who post > 2 times in a month): 0.75%
- Average Mildly Active (members who post one or two times): 2.26%
- Average of members who post anything at all: 3.01%
I chose > 2 to mean ‘active’ because it is so minimal that I couldn’t go to 1 or 2 and anything higher saw a big drop off in that percentage. I provide two months of data to see if it’s stable (and it is).
It’s possible that it’s closer to the 90–9–1 if you think that the 97 is really split itself between lurkers and join-only people. If a typical group is 50% join-only (join but then completely ignore it) and we remove those join-only people, the percentages become something like 94–5–1.”
2. Community of Practice Data
Lee and I also analyzed data from the SIKM Leaders Community. Here is Lee’s summary:
“If I take the months Jan 2015-April 2016 and average the percent active and percent very active, that works out to 2.1% and 1.77%, respectively. Which is then more like 96–2–2.”
3. LinkedIn Data
One more set of data comes from my LinkedIn posts. Looking at the number of views, likes, and comments for each post, out of 6,229 followers, I see numbers like:
- 308, 52, 5
- 589, 72, 11
- 557, 60, 7
- 635, 68, 14
- 416, 27, 4
- 476, 45, 9
- 847, 70, 5
- 805, 119, 19
- 384, 33, 2
- 475, 69, 3
The ratios form a pattern of 5–10% of followers who read a post, .5–1% of followers who like a post, and .05-.1% of followers who comment. This is a familiar power law distribution, in this case 90% don’t read, 10% read, 1% like, and .1% comment.
4. Join-Only Members
Expanding on 90–9–1, it would be useful to add a fourth number representing the people who join communities, but don’t pay any attention to what is going on. This makes a distinction in the lurkers between learners and join-only members.
Learners pay attention to the community’s discussions, shared content, and events. Join-only members may have good intentions, but they end up being essentially the same as non-members, because they receive no benefits from being members.
There is nothing wrong with the 90% not posting, as long as they read, listen, pay attention, etc. But if they don’t, then they are not getting value from the group, and the organization misses out on their personal development and/or their contributions to the other members.
5. Is the 90–9–1 rule of thumb invalid?
From Is the 90–9–1 Rule for Online Community Engagement Dead? by Paul Schneider:
“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”
I haven’t seen this for communities. In fact, I see more evidence of 95–4–1, even in groups focused on knowledge management and collaboration. And some think that the number of inactive members should be even higher.
From Guesstimating the accuracy of the 90/9/1 rule in Reddit:
“I think it’s clear that the reality could well be more like the 98/1.9/0.1 rule, instead of what the accepted reality has been so far.”
From Three Community Myths Busted by Ted McEnroe of the Community Roundtable:
“Myth: The 90–9–1 rule — just one percent of members of a community are truly engaged, while 90 percent lurk.
Reality: More like 55–25–20.
Communities are complex organisms, and really, no one “rule” will define engagement percentages for every community. But consistently over the past three years, our research has found that engagement levels in communities are consistently higher than the old assumption.
That’s not to say there aren’t big, external (usually) communities that aren’t 90–9–1. And whether you count inactive but registered members, how you set up the community permissions and other items can have an effect on your numbers. But when you normalize for members who log into a community during a given month — you’re more likely to find that a sizable minority are posting, commenting, creating, liking, sharing and collaborating.”
Lee Romero commented:
“A big caveat in that last line. That is like the join-only member. So if you ignore the fact that, say, 50% (or whatever the percent is) just join and then ignore everything, you can argue that the percentages are higher, sure. You could extend that point, though. ‘When you normalize for members who post in a community during a given month, we find that 100% of members post at least once a month.’ A true statement but it doesn’t necessarily change anything.”
From The 2013 State of Community Management by the Community Roundtable:
“Demystifying the myth of the 90–9–1 rule
The biggest finding of our research was quantifying just how much active community management correlates with high levels of engagement. Most community management professionals are familiar with the 90–9–1 rule of online engagement and some communities do track very well to that engagement pyramid. But many question this rule because it can be unreliable or an underestimate.
What we found in our survey was striking — an average engagement profile of 55–30–15 — wildly different than the common rule of thumb. More surprising, the average of the most engaged communities reported more creators than lurkers — at 17–57–26. For best-in-class communities there are more content creators than there are lurkers”
Note that this data is taken from a survey, which is different from directly measuring data in actual communities. I prefer to rely on empirical data, rather than survey data.
6. How can the participation rate be increased?
If you can generate a higher percentage of active community members, that’s great. For internal communities, this usually depends on factors outside the direct control of the community manager, such as senior leaders who are active and get others to follow their example.
From The 90–9–1 Rule for Participation Inequality in Social Media and Online Communitiesby Jakob Nielsen:
“How to Overcome Participation Inequality
The first step to dealing with participation inequality is to recognize that it will always be with us. It’s existed in every online community and multi-user service that has ever been studied.
Your only real choice here is in how you shape the inequality curve’s angle. Are you going to have the ‘usual’ 90–9–1 distribution, or the more radical 99–1–0.1 distribution common in some social websites? Can you achieve a more equitable distribution of, say, 80–16–4? (That is, only 80% lurkers, with 16% contributing some and 4% contributing the most.)”
7. Is there any downside to increasing the percentage of active participants?
What would happen if 90% of community members posted regularly? In a small group, that might be okay. But in a large community, the volume and frequency of posts would likely become a problem.
More frequent posts from more people can result in more noise, and can cause members to stop paying attention. So those who lament that only 10% of a community are active should consider the possible negative consequences of a dramatic increase.
If people will just pay attention, read, and learn, that is valuable. And based on my experience, that will be the most we can expect from most members.
8. Zipf’s Law
From Describing the distribution of engagement in an Internet support group by post frequency: A comparison of the 90–9–1 Principle and Zipf’s Law by Bradley Carron-Arthur, John Cunningham, and Kathleen Griffiths:
“The current analysis broadly replicated the findings of van Mierlo (2014), that the top 1% of registered members contribute the vast majority of posts, the next 9% a minority and the last 90% very few. Thus, the 90–9–1 principle appears to provide a reliable means of broadly categorising participant contributions in a DHSN. However, the graph in Fig. 2 and the associated best fitting power curve provide an alternate and more precise means of describing the distribution.
In fact, the distribution in Fig. 2 adheres to Zipf’s law — that the frequency of posts made by a member is inversely proportional to their rank in frequency. This is a widely observed phenomenon spanning areas such as linguistics, populations, income and internet traffic (Newman, 2006; Adamic and Huberman, 2002). This model gives a more nuanced image of the distribution. It shows a gradual reduction in contributions rather than a quantum leap at the boundary between superusers and contributors as the 90–9–1 principle implies.”
9. Does 90–9–1 even matter?
Another query I received: “What share of your community members actively post (vs. consume/lurk)? And at what rate do they post?” This prompted two opposing replies:
From Mark Diller:
“1–9–90 matters. Communications departments are often Yammer advocates within the organization, and those departments have established methods of communication: intranet, email, face-to-face, etc. When you add Yammer to the mix, the first question you start fielding is how effective this new social network is in raising awareness of a given topic. If all you’re doing is reporting the ‘1’ and the ‘9,’ it looks like Yammer pretty much sucks by comparison with the other vehicles. The ‘90’ is vague, hypothetical, and largely unverifiable, but it’s an important element of estimating the overall reach of your network.”
From Steve Nguyen:
“Does 90/9/1 really matter? Is the goal really to convert the 90s to the 1s? Sure, the benefit comes when there are a more contributors. But contributing about what? I think that’s the more important aspect to focus on. If we can get our communities focused on contributing relevant content, those numbers kind of work themselves out.
As an example, I have no idea what the 90/9/1 ratio is for our Microsoft network or this Office 365 community. But I’m confident enough in the existing contributors and contributions that I can find the information I need when I need it. So from that standpoint, does 90/9/1 matter? Not really. As a participant in the network, I trust the people in it.”
- Presentation version of this article
- Twitter Discussion and subsequent SIKM Discussion
- Power Laws, Weblogs, and Inequality by Clay Shirky
- 90:9:1 — the odd ratio that technology keeps creating by Charles Arthur
- Is it true that over 50% of all edits on Wikipedia are done by the 524 most active users? in Quora
- 1% rule in Wikipedia
- Community Management: The 90–9–1 Rule is Dead by Sam Fiorella
- Ripples of influence in a CoP, moving through the 90–9–1 rule by Ewen Le Borgne
- The 1% rule and why it still matters by Ramy Khuffash
- Where is Everybody? The 90–9–1 Rule Explains Where Your Blog Audience Hides by Dani Finklestien
- Is the 1% rule dead? The BBC thinks so, but it’s wrong by Bobbie Johnson
- The 90–9–1 Rule, Forums, and Social Media by Adriaan Bloem
- The 90–9–1 Rule in Reality by Michael Wu
- The Economics of 90–9–1: The Lorenz Curve by Michael Wu
- The Economics of 90–9–1: The Gini Coefficient (with Cross Sectional Analyses)
- Is the 1% Rule Still Relevant? by Kevin Spidel
- The 90–9–1 Rule: Dead, Different, or a Distraction? by Crystal Coleman
- Mapping Power Law Distributions in Digital Health Social Networks: Methods, Interpretations, and Practical Implications by Trevor van Mierlo, Douglas Hyatt, and Andrew T Ching
- Community by the Numbers, Part III: Power Laws by Christopher Allen
- The 90–9–1 Collaboration Paradox: Org’s Should Aim To Reverse It by Dan Pontefract
- Leveraging the 1% rule by John Stepper
- Lurkers are learners: new approaches to understanding participation by Catherine Shinners
- Why Lurkers Lurk by Blair Nonnecke and Jenny Preece
- Ten Principles for Managing 90–9–1 for online community engagement by Crispin Butteriss
Do you have actual community posting data, not from a survey? If so, please share it!