Very good idea, and very consumable reference. I hope it spreads.
Though I wonder if the 'editor' label is the best choice for the 9% section. In my experience the 90-9-1 rule applies to the degree of contributions, not the type. While it seems kind of like a good idea to try to marry the current participation models (Forrester's Social Technographics, Gartner's Four Levels, etc.) with the 1% rule, I'm not sure it really works that way. A key point to note is that participation types in those models usually overlap (the numbers add up to more than a 100%). And in some of those models 'Creators' etc. can be significantly more than 1%.
I guess any label you choose here is going to have difficulties, but I wonder if trying to classify participation at the same time could be misleading. But still a very useful reference for those new to the topic. Thanks for putting this together!
Jake McKee
· 11 months ago
Seesmic video reply from Disqus.
Jake McKee
· 11 months ago
Seesmic video reply from Disqus.
EricFriedman
· 11 months ago
Hi Jake,
This is great - and will hopefully become a go-to for people looking to explain the principle.
Scott Moore
· 11 months ago
Jake, last night at the SF Community Meetup, folks were mentioning push-back they get from executives about 10% participation. If you haven't seen it, I found this whitepaper from Rubicon Consulting pretty interesting in helping show that while 10% may be generating/editing the bulk of the content, their influence is much larger (the other 90%).
I agree with Scott D that trying to merge models will lead to difficulties. That's why I love this quote: "All models are wrong, but some are useful." - George E.P. Box, Professor Emeritus, College of Engineering, University of Wisconsin-Madison
It really just means we have multiple tools with which to describe behavior (participation degree and participation type). Thanks Jake for collecting one of these tools into an easily digestible format.
Jake McKee
· 11 months ago
Seesmic video reply from Disqus.
Jeremy
· 11 months ago
Jake, I think that you are a little confused in your terminology, and are using "early adopters" and "early majority" as synonymous buzzwords. They are actually specific phases in the Adoption Lifecycle, or Roger's Bell Curve, that is quite standard in marketing pedagogy.
Jake, you broke the pyramid! But seriously, Neilson's model is very hard to improve upon.
Here's just one way in which the "creator/editor" hierarchy breaks down: As you note in your audio comment, one way to think about creator v. editor is that the creator starts the conversation and the editor refines and improves it. But it's exactly the opposite of what happens in most forum-based communities. In most forums, the 1% hardly ever start a conversation. They reply, but the don't start. The 9% are the ones who start the conversations. It's almost universal for communities larger than a few hundred people. So your top two labels could be reversed.
But actually you've munged Neilson's hierarchy together with a Forrester/Gartner-style participation ladder, and I'd argue that they do different things.
Still, an excellent way to kick off the conversation ...
Joe
Jake McKee
· 11 months ago
That's a relevant point Joe. Let me chew on that a bit more.
As far as Forrester, I don't mean to imply that I'm swiping the meaning by using the same words.
Brinstar
· 11 months ago
I was wondering whether you think this principle applies to all social media spaces. Specifically, I'm wondering about online computer game (MMORPGs like World of Warcraft) fan communities.
Jake McKee
· 11 months ago
I do believe that the principle applies, although the definitions of each stage of the 90, 9, and 1 percentages probably differ from space to space.
In the MMORPG there are varying levels of activity, even if "activity" is, by default a type of "creation". While absolutely everyone in an MMORPG might do certain types of creation activities as a simple by-product of playing the game at all, there's a much smaller percentage of people doing things like organizing guilds, creating out-of-game web sites, etc.
Joe
· 11 months ago
This site is a great idea, Jake - I don't think this formula is as widely known or as well understood as it should be.
I'll join the debate around labels for each of the three segments of the pyramid: Nielsen's labels are really very simple and very accurate -- heavy contributor, intermittent contributor, and lurker. In his model, the main difference between the top two segments is one of degree, not kind.
Are there differences in kind between what the 1% do and what the 9% do? Yes - but in my experience the difference is exactly the opposite of what you describe in your response to Scott. In your model, the Creator "sits down with a blank piece of paper," and the Editor "appends something to something someone else has started." In fact, in the communities I know, the 1% almost never initiate the conversation -- that work is done by the 9%. The contributions of the 1% are overwhelmingly in the form of replies to what others have created. I'd be very surprised if this isn't your experience as well. So your labels should actually be reversed -- though I'd prefer a return to Nielsen's more accurate ones.
The concept of "editing" in communities is actually a very interesting one. Only in wiki-based communities is editing a privilege generally extended to all members. The average user can't edit a blog post or comment (other than their own), and can't edit a post in a forum (other than their own). If you extend the definition of editing to include "appending" or "replying," then you end up with the problem I noted above -- the 9% mostly author and don't "edit."
My last observation relates to the label "audience." I've never been a huge fan of "lurker," for all the obvious reasons. But "audience" belongs more to a world of 1.0 (or even 0.0) than to Web 2.0. I sometimes use the word "target audience" when I'm talking to the companies about the communities they could create, but once a community is up and running it doesn't feel accurate to describe non-posters as if they are watching television or attending a concert. Among other things, it's desireable to have a label for this segment that suggests the potential of lurkers to "de-lurk" -- whereas audience members can't really step up on stage.
One thing we don't talk about enough re 90-9-1 is that it is a point-in-time snapshot and not a longitudinal view. The boundaries between the slices is fluid. The 10% who participate this month are not identical to the 10% who participated last month. So every time I cite this statistic I make sure I qualify it by stressing the fact that it does not mean that only 10% of your target audience will ever participate. It means simply that only 10% will participate at any one time.
Miles Sims
· 11 months ago
Very nice! This is something we cover with clients all the time and is a great tool to help people understand that their are different types of people using their community and creating a different level of relationship with each group.
Mike Rowland
· 11 months ago
Hi Jake,
I'm going to play the devil's advocate on this. We've seen this ratio thrown out as the general rule of thumb for many years. In fact, we used it back in 2000 to educate new clients at Participate.com to set expectations. However, in the real world as you state above the ratios are very different. We've seen a support community where the ratio would be more like 15-20-65 and a social network where the ratio would be more like .001-3-96.009. In B2B communities, we see very different ratios. Depending upon the tool used (forums vs. blog vs. wiki) we see very different ratios.
The general ratio also does not take into account the role of active facilitators in the community. Without them, the ratio falls significantly. Facilitators are not the 1 or the 9%, they are outside the group as they are employed by the site. Too often, organizations will focus on the technology not the social dynamics and then wonder why they cannot get to the 1-9-90 ratios they've been told naturally occur.
This measure is useful as a starting point for discussion however. But I would not use this to set expectations or objectives for clients. In my opinion, we need to move away from sweeping generalizations and present experiential examples to truly help people understand the power of social media and how their audience may use it. Community/social media has been around for over 10 years. Is this really the best we can provide? Perhaps, but I would hope that others (especially the firms in the Community Management business) could help us with a better set of expectations that are based on reality, not a generalized perception.
I'll jump off my soapbox now... ;-)
Mike
Jake McKee
· 11 months ago
Great thoughts, Mike! Thanks for sharing.
"But I would not use this to set expectations or objectives for clients."
Absolutely. In fact, I'd go so far as to say that using a model, any model, to set client expectations leaves your client in bad shape. This site, and hell the rule itself, is meant to be a teaching tool; one that says "there are a small number of people who do a majority of the work". As Joe points out, not only may the percentages adjust by tool/client/project/community, but the roles may actually flip-flop.
Yes, we've been doing this a while but the rest of the world is just now catching up. While we might want get to the point of using more sophisticated models and discussions, sadly our clients and colleagues aren't quite there. Yet.
Joe Cothrel
· 11 months ago
Thanks, Jake -- one of my favorite topics!
I'm with Scott and Scott on the labels -- Nielsen's are clear and accurate: heavy contributor, intermittent contributor, lurker. Creator, editor, and audience are clear but not accurate.
Here's the trouble: in your response to Scottd, you described Creators as the people who "start with a blank piece of paper," and Editors as those "who append something to something someone else has started." However, in communities I know -- those where users provide help and advice to other users -- it's exactly the opposite. The 1% almost never start a conversation. It's the 9% who ask, and the 1% who reply.
I also wonder about the label "audience." I'm no fan of "lurker" -- it's always seemed pejorative to me -- but "audience" suggests a relation similar to watching television or attending a concert. Lurker at least implies the possibility of "delurking." Audience seems more 1.0 -- or even 0.0 -- than 2.0.
So I think we'd have a more productive conversation if we began with Nielsen's model and explored that more thoroughly. There are a lot of interesting questions around that, including how the model applies to communities of different size, composition, purpose, or mode of interaction.
Jake McKee
· 11 months ago
(D'oh! Sorry for encouraging you to double post when we thought your comment was eaten. But it was different and good, so I'm leaving it up! )
Adam Bow
· 11 months ago
After reading about this rule I immediately began to think about certain segments and groups and how much this rule would vary based on those various groups. This is a very broad rule based off all data available. It would be fascinating to examine various interest groups. For instance, looking at moms and their growing numbers of adoption, interaction, and participation on the web this rule has to most certainly change. More importantly strategies to change the dynamics of the rule (methods of increasing the 1%) would change and rely heavily among that target group culture.
A quick look at stats among the mom group: YouTube reaches 21.1 million moms, 42% of moms use photo sharing sites, over 60% write product reviews, 44% write on bulletin boards, 80% of moms read blogs, and most importantly 78% of mom blogs are about products.
The top 3 activities for Gen Y moms are reading blogs, participating in online community of moms, and creating and sharing their own video. Top 3 activities of Gen X moms are using photo sites, rating and reviewing products, and shopping.
Here is a group who's largest online activity is rating and reviewing products, participating in online communities and blogs, and creating and sharing their own content. I would argue that by creating tools (applications/environments) that connect to their behavior online will substantially increase that 1%.
abbeyrae
· 11 months ago
Love the new site! this is a great find!
Shaun Abrahamson
· 8 months ago
Sorry I am late to the comments. Was hoping to find more 80-20 rules to keep things simple. but alas.
From the conversation, I understand that this is meant as a teaching aid when engaging people who are just starting to think about this. IN fact, I found my way here by asking someone at Spinspotter where they got the rule from (they referenced it in their pitch).
I am wondering if you see emerging trends. For example, people who grow up with online communities tend to contribute more. Or people in countries like China (online population seems to skew younger). Or specific types of community (for example product ideas versus comments on a blog).
Finally, might it be that in some cases the value is that someone has already done the work. i.e. if I show up at an answers site, the answer might be there. Therefore no need to interact. So the 90 is just a byproduct of how many people it took to answer, versus the total traffic visiting the answer.
Further, are you saying that the same people play the same roles over time? This seems to make some sense when I have looked at sites where it is possible to observe level of participation via points, for example.
Thanks for a great resource and your selection of Creative Commons license.
Westy | PHOTONOMY
· 8 months ago
Excellent post and site.. i will be citing it in my dissertation whyNOTparticipation
i am final year student at Bournemouth Media School and think this is spot on with in the view of particpation being an active contribution.
LaurelPapworth
· 6 months ago
The 90:9:1 Rule was true when Dr Jakob Nielsen came up with it. Mostly because only a few years before we were the 100% passive community. We have slowly learnt that it's ok to contribute publicly, to strangers our responses and our creativity.
Most communities these days are much higher - TripAdvisor has a 30% creator community. Twitter has a much lower number of lurkers than respondents.
I'm sorry but this is an outdated model that damages social network strategies and gives companies excuses for enabling comment only and not collaboration. And they excuse poor numbers of comments and participation with "oh never mind, only 1% were ever going to create content around our ideas anyway, and only 9% would've ever left a comment". Let's not propagate this archaic principle, hmmm?
socialscience101
· 4 months ago
Where does this data come from?
I'd like to see the research, before I can agree with the principle.
All Comments
Though I wonder if the 'editor' label is the best choice for the 9% section. In my experience the 90-9-1 rule applies to the degree of contributions, not the type. While it seems kind of like a good idea to try to marry the current participation models (Forrester's Social Technographics, Gartner's Four Levels, etc.) with the 1% rule, I'm not sure it really works that way. A key point to note is that participation types in those models usually overlap (the numbers add up to more than a 100%). And in some of those models 'Creators' etc. can be significantly more than 1%.
I guess any label you choose here is going to have difficulties, but I wonder if trying to classify participation at the same time could be misleading. But still a very useful reference for those new to the topic. Thanks for putting this together!
This is great - and will hopefully become a go-to for people looking to explain the principle.
http://tinyurl.com/57p323
I agree with Scott D that trying to merge models will lead to difficulties. That's why I love this quote: "All models are wrong, but some are useful." - George E.P. Box, Professor Emeritus, College of Engineering, University of Wisconsin-Madison
It really just means we have multiple tools with which to describe behavior (participation degree and participation type). Thanks Jake for collecting one of these tools into an easily digestible format.
I think that you are a little confused in your terminology, and are using "early adopters" and "early majority" as synonymous buzzwords. They are actually specific phases in the Adoption Lifecycle, or Roger's Bell Curve, that is quite standard in marketing pedagogy.
http://www.valuebasedmanagement.net/methods_rog...
Regards
Here's just one way in which the "creator/editor" hierarchy breaks down: As you note in your audio comment, one way to think about creator v. editor is that the creator starts the conversation and the editor refines and improves it. But it's exactly the opposite of what happens in most forum-based communities. In most forums, the 1% hardly ever start a conversation. They reply, but the don't start. The 9% are the ones who start the conversations. It's almost universal for communities larger than a few hundred people. So your top two labels could be reversed.
But actually you've munged Neilson's hierarchy together with a Forrester/Gartner-style participation ladder, and I'd argue that they do different things.
Still, an excellent way to kick off the conversation ...
Joe
As far as Forrester, I don't mean to imply that I'm swiping the meaning by using the same words.
In the MMORPG there are varying levels of activity, even if "activity" is, by default a type of "creation". While absolutely everyone in an MMORPG might do certain types of creation activities as a simple by-product of playing the game at all, there's a much smaller percentage of people doing things like organizing guilds, creating out-of-game web sites, etc.
I'll join the debate around labels for each of the three segments of the pyramid: Nielsen's labels are really very simple and very accurate -- heavy contributor, intermittent contributor, and lurker. In his model, the main difference between the top two segments is one of degree, not kind.
Are there differences in kind between what the 1% do and what the 9% do? Yes - but in my experience the difference is exactly the opposite of what you describe in your response to Scott. In your model, the Creator "sits down with a blank piece of paper," and the Editor "appends something to something someone else has started." In fact, in the communities I know, the 1% almost never initiate the conversation -- that work is done by the 9%. The contributions of the 1% are overwhelmingly in the form of replies to what others have created. I'd be very surprised if this isn't your experience as well. So your labels should actually be reversed -- though I'd prefer a return to Nielsen's more accurate ones.
The concept of "editing" in communities is actually a very interesting one. Only in wiki-based communities is editing a privilege generally extended to all members. The average user can't edit a blog post or comment (other than their own), and can't edit a post in a forum (other than their own). If you extend the definition of editing to include "appending" or "replying," then you end up with the problem I noted above -- the 9% mostly author and don't "edit."
My last observation relates to the label "audience." I've never been a huge fan of "lurker," for all the obvious reasons. But "audience" belongs more to a world of 1.0 (or even 0.0) than to Web 2.0. I sometimes use the word "target audience" when I'm talking to the companies about the communities they could create, but once a community is up and running it doesn't feel accurate to describe non-posters as if they are watching television or attending a concert. Among other things, it's desireable to have a label for this segment that suggests the potential of lurkers to "de-lurk" -- whereas audience members can't really step up on stage.
One thing we don't talk about enough re 90-9-1 is that it is a point-in-time snapshot and not a longitudinal view. The boundaries between the slices is fluid. The 10% who participate this month are not identical to the 10% who participated last month. So every time I cite this statistic I make sure I qualify it by stressing the fact that it does not mean that only 10% of your target audience will ever participate. It means simply that only 10% will participate at any one time.
I'm going to play the devil's advocate on this. We've seen this ratio thrown out as the general rule of thumb for many years. In fact, we used it back in 2000 to educate new clients at Participate.com to set expectations. However, in the real world as you state above the ratios are very different. We've seen a support community where the ratio would be more like 15-20-65 and a social network where the ratio would be more like .001-3-96.009. In B2B communities, we see very different ratios. Depending upon the tool used (forums vs. blog vs. wiki) we see very different ratios.
The general ratio also does not take into account the role of active facilitators in the community. Without them, the ratio falls significantly. Facilitators are not the 1 or the 9%, they are outside the group as they are employed by the site. Too often, organizations will focus on the technology not the social dynamics and then wonder why they cannot get to the 1-9-90 ratios they've been told naturally occur.
This measure is useful as a starting point for discussion however. But I would not use this to set expectations or objectives for clients. In my opinion, we need to move away from sweeping generalizations and present experiential examples to truly help people understand the power of social media and how their audience may use it. Community/social media has been around for over 10 years. Is this really the best we can provide? Perhaps, but I would hope that others (especially the firms in the Community Management business) could help us with a better set of expectations that are based on reality, not a generalized perception.
I'll jump off my soapbox now... ;-)
Mike
"But I would not use this to set expectations or objectives for clients."
Absolutely. In fact, I'd go so far as to say that using a model, any model, to set client expectations leaves your client in bad shape. This site, and hell the rule itself, is meant to be a teaching tool; one that says "there are a small number of people who do a majority of the work". As Joe points out, not only may the percentages adjust by tool/client/project/community, but the roles may actually flip-flop.
Yes, we've been doing this a while but the rest of the world is just now catching up. While we might want get to the point of using more sophisticated models and discussions, sadly our clients and colleagues aren't quite there. Yet.
I'm with Scott and Scott on the labels -- Nielsen's are clear and accurate: heavy contributor, intermittent contributor, lurker. Creator, editor, and audience are clear but not accurate.
Here's the trouble: in your response to Scottd, you described Creators as the people who "start with a blank piece of paper," and Editors as those "who append something to something someone else has started." However, in communities I know -- those where users provide help and advice to other users -- it's exactly the opposite. The 1% almost never start a conversation. It's the 9% who ask, and the 1% who reply.
I also wonder about the label "audience." I'm no fan of "lurker" -- it's always seemed pejorative to me -- but "audience" suggests a relation similar to watching television or attending a concert. Lurker at least implies the possibility of "delurking." Audience seems more 1.0 -- or even 0.0 -- than 2.0.
So I think we'd have a more productive conversation if we began with Nielsen's model and explored that more thoroughly. There are a lot of interesting questions around that, including how the model applies to communities of different size, composition, purpose, or mode of interaction.
A quick look at stats among the mom group: YouTube reaches 21.1 million moms, 42% of moms use photo sharing sites, over 60% write product reviews, 44% write on bulletin boards, 80% of moms read blogs, and most importantly 78% of mom blogs are about products.
The top 3 activities for Gen Y moms are reading blogs, participating in online community of moms, and creating and sharing their own video. Top 3 activities of Gen X moms are using photo sites, rating and reviewing products, and shopping.
Here is a group who's largest online activity is rating and reviewing products, participating in online communities and blogs, and creating and sharing their own content. I would argue that by creating tools (applications/environments) that connect to their behavior online will substantially increase that 1%.
From the conversation, I understand that this is meant as a teaching aid when engaging people who are just starting to think about this. IN fact, I found my way here by asking someone at Spinspotter where they got the rule from (they referenced it in their pitch).
I am wondering if you see emerging trends. For example, people who grow up with online communities tend to contribute more. Or people in countries like China (online population seems to skew younger). Or specific types of community (for example product ideas versus comments on a blog).
Finally, might it be that in some cases the value is that someone has already done the work. i.e. if I show up at an answers site, the answer might be there. Therefore no need to interact. So the 90 is just a byproduct of how many people it took to answer, versus the total traffic visiting the answer.
Further, are you saying that the same people play the same roles over time? This seems to make some sense when I have looked at sites where it is possible to observe level of participation via points, for example.
Thanks for a great resource and your selection of Creative Commons license.
i am final year student at Bournemouth Media School and think this is spot on with in the view of particpation being an active contribution.
Most communities these days are much higher - TripAdvisor has a 30% creator community. Twitter has a much lower number of lurkers than respondents.
I'm sorry but this is an outdated model that damages social network strategies and gives companies excuses for enabling comment only and not collaboration. And they excuse poor numbers of comments and participation with "oh never mind, only 1% were ever going to create content around our ideas anyway, and only 9% would've ever left a comment". Let's not propagate this archaic principle, hmmm?
I'd like to see the research, before I can agree with the principle.