In short
The 90-9-1 rule from 2006 says most users lurk while a tiny group does most of the posting. Later research and Tryane's Viva Engage Benchmark 2026 confirm the concentration, but the split varies sharply by company size, so set contribution targets for your own segment instead of quoting the rule.
Key takeaways
- Jakob Nielsen's 2006 rule: 90% of community users lurk, 9% contribute a little and 1% do most of the posting.
- A 2014 study of four digital health networks found Superusers produced 74.7% of all content, confirming the concentration.
- In Tryane's Viva Engage Benchmark 2026, contributors range from 10.2% under 2,500 staff to 2.4% at 25,000 to 50,000 employees.
- You cannot remove participation inequality, only shape it: lower the barrier and make participation a side effect.
- Microsoft 365 admin reports count each user's posts, reads and likes, so readers who never post are measurable.
What is the 90-9-1 rule?
In 2006, Jakob Nielsen of the Nielsen Norman Group condensed decades of community research into a rule of thumb that still shapes how we talk about social platforms today. In most online communities, he observed, 90% of users are lurkers who read or observe but never contribute, 9% contribute from time to time, and 1% account for almost all the action.[1] Nielsen called this phenomenon participation inequality, and if you manage a Viva Engage network, you have probably felt it: a handful of familiar names fill your communities while most colleagues stay silent.
- 90% of users are lurkers: they read or observe, but do not contribute.
- 9% of users contribute from time to time, but other priorities dominate their time.
- 1% of users participate a lot and account for most contributions, often posting within minutes of the events they comment on.
Where the imbalance comes from
Participation inequality was first studied in depth by Will Hill in the early 1990s at Bell Communications Research, and Nielsen found the same pattern across Usenet newsgroups, mailing lists and internal company discussion boards. When you plot activity per user, you get a Zipf curve: a tiny elite contributes massively, and a long tail contributes almost nothing. The imbalance is not a quirk of one platform. Nielsen showed that blogs follow an even steeper 95-5-0.1 split, and Wikipedia is steeper still, with about 0.2% of visitors doing nearly all the editing.[1]
For anyone who runs an internal network, Nielsen's sharpest warning is about interpretation, not measurement. In Whittaker et al.'s classic Usenet study, a randomly selected post was equally likely to come from one of 19,000 high-frequency contributors or one of 580,000 low-frequency contributors.[1] Reading the community through its posts therefore gives you a biased understanding of it, because you never hear from the silent majority. Translate that to your own Viva Engage feed: the colleagues who comment the loudest are not a representative sample of how your workforce thinks.
Does the 90-9-1 rule still hold in practice?
The rule is a heuristic from 2006, so it is fair to ask whether later data confirms it. The most cited test came in 2014, when researchers analysed four long-standing digital health social networks and published their results in the Journal of Medical Internet Research. The concentration they found was real, and it was even harsher than 90-9-1 suggests.[2]
- 63,990 actors created 578,349 posts across the four networks during the study period.[2]
- Fewer than 25% of actors ever made a single post.[2]
- Superusers produced a weighted average of 74.7% of all content.[2]
- Contributors accounted for 24.0% of content, and Lurkers for just 1.3%.[2]
What the study confirms is the core of Nielsen's argument: in any community, content stays concentrated in a small group, and the people who create it are not typical of the people who benefit from it. What it qualifies is the exact split. The authors concluded that the 1% rule was consistent across their four networks, but the shares of content produced by Lurkers, Contributors and Superusers varied from network to network, so the boundaries between the three groups fall in different places everywhere. The rule describes a shape, not a law.[2]
For internal communication teams, the practical implication is straightforward. In your own Viva Engage network, a handful of posters will always dominate what you see in the feed. That is expected behaviour, not a sign that your content has failed, and it is exactly why you should not judge reach by what the feed shows you.
Who posts on Viva Engage, by company size?
So what does participation inequality look like on Viva Engage specifically? Tryane's Viva Engage Benchmark 2026 analyses the activity of more than 3.5 million users across customer networks and breaks down the share of users who read, like or contribute in each period, by company size. The pattern is striking: contribution falls sharply as headcount grows.
| Company size | Readers | Likers | Contributors |
|---|---|---|---|
| Fewer than 2,500 employees | 49% | 22.4% | 10.2% |
| 2,500 to 5,000 employees | 61% | 11.1% | 4.1% |
| 5,000 to 10,000 employees | 52% | 10.8% | 6.7% |
| 10,000 to 25,000 employees | 58% | 6.1% | 3.3% |
| 25,000 to 50,000 employees | 55% | 4.8% | 2.4% |
| More than 50,000 employees | 43% | 6.3% | 5.2% |
In organisations with fewer than 2,500 employees, 22.4% of users like posts and 10.2% contribute, the highest shares of any size band. At 25,000 to 50,000 employees, those shares drop to 4.8% and 2.4%, the lowest of any band. Readers, by contrast, are the largest group everywhere, peaking at 61% in the 2,500 to 5,000 band and never falling below 43%. The benchmark's own commentary draws the conclusion that matters for you: high readership with low interaction highlights the need for better activation strategies to turn silent readers into active participants. To compare your own network by industry and company size, download the full Viva Engage Benchmark 2026.
One caution before you quote these numbers upward: they are shares of users by behaviour, not a verification of 90-9-1. A network where a tenth of users contribute is not 'beating' the rule, and one where only a fortieth contribute is not 'failing' it. The benchmark measures different things than Nielsen's heuristic did, which is why the next sections focus on what you can actually do with your own segment's numbers.
Can you change participation inequality?
Nielsen's answer is honest: you cannot remove participation inequality, you can only shape it. Every attempt to eliminate lurking entirely has failed, because reading without posting is how most people legitimately use social tools. Your only real choice is in how you shape the inequality curve's angle: whether you keep the usual 90-9-1 distribution or move towards something more equitable.[1]
- Lower the barrier to contribution. Nielsen's examples include making it effortless to act, with one-click responses rather than blank-page moments. On Viva Engage, that means polls, quick reactions and questions that need a one-line answer, not essay prompts.
- Make participation a side effect. When people can contribute as part of something they already do, such as answering a colleague's question in the community where they already work, contribution stops competing with their real workload.
- Reward contributors, but do not over-reward the elite. Recognise new voices and first-time commenters explicitly, so the same ten names do not collect all the visibility and the perceived cost of joining stays low.
Each lever maps directly to work you already do as an IC manager. A poll in an announcement community is a lower barrier than a comment thread. A question posted by a leader that invites short answers invites the 9% to move, not just the 1%. And structuring communities around real teams and topics, rather than catch-all groups, gives silent readers an obvious, low-risk place to start. We cover the community side of this in more depth in our guide to creating relevant Viva Engage communities.
Above all, treat lurking as normal, not as a failure of your communication. The benchmark data above shows that even in the most active networks, most users read. A colleague who reads every post in your leadership community is reached, even if they never click like.
How do you measure readers who never post?
If lurking is normal, your measurement needs to count it. The good news is that reading is measurable on Viva Engage without any third-party tool. The Viva Engage Activity report in the Microsoft 365 admin center shows the number of unique users who post, like or read a message, with trends over the last 7, 30, 90 or 180 days.[3]
- Posted: the number of messages the user posted in the period you select.
- Read: the number of conversations the user read in that period.
- Liked: the number of messages the user liked in that period.
- Last activity date: the last date the user posted, read or liked a message.
- Export: you can export the full per-user table to an Excel CSV file for further sorting and filtering.
That export is where the reader majority becomes visible: sort by the Read column and the silent 90% appears, with real activity dates. But the native report has limits for IC teams. By default, user-specific information like usernames and display names is hidden in usage reports and must be made visible by an admin, and if you select a particular day in the report, the table shows data for up to 28 days from the current date, not the date the report generated.[4] There is also no department, country or audience-segment view. For a fuller picture of what the built-in reports do and where they stop, see our guide to Viva Engage analytics.
The next analytical layer is the one that changes decisions: seeing which departments read but never react, or which countries like but never comment. That is the difference between knowing that a majority of your organisation reads, and knowing which specific audience you are failing to activate. It is also where dedicated analytics picks up, because segment-level reader data is what turns a participation report into a targeting decision.
You now have everything you need to set realistic goals. Use 90-9-1 as context when you explain engagement to leadership: it explains why a small group dominates the feed, and why posts alone are a poor measure of reach. But set your targets from your own size band, because the benchmark shows contribution varies by a factor of four across company sizes, from about 2.4% of users contributing at 25,000 to 50,000 staff to 10.2% under 2,500.
- Anchor your baseline in your size band, not in 90-9-1. A contributor ambition of 10.2% is realistic under 2,500 employees and unrealistic at 25,000 to 50,000, where 2.4% is the benchmark average.
- Report readership as your reach signal. Readers range from 43% to 61% depending on band, which makes readership a far more stable KPI than posts or comments, and the one that reflects how most employees actually use the network.
- Move likers first. Liker shares are consistently between roughly 1.2 and 2.7 times contributor shares, so low-effort formats such as polls and reactions are your shortest path to activating silent readers before you push for comments.
- Review against the same period each quarter. Compare like with like, quarter on quarter, so trends reflect your activation work rather than seasonal noise.
Choosing which numbers to report is itself a strategic decision, and it is worth doing deliberately rather than defaulting to whatever the platform shows first. Our guide on what to measure in internal communication analytics walks through the four questions, distribution, reach, engagement and outcome, that turn raw platform data into a story leadership can act on. The goal is not a prettier dashboard. It is the ability to make data-driven decisions about where to invest your team's limited time.
Everything above rests on one capability: knowing who reads, who likes and who posts in your network, segment by segment. Analytics for Viva Engage gives you that view. It monitors conversations across your communities, identifies the content that engages most, and shows you readers, likers and contributors in your own network, so you can benchmark contribution per segment instead of guessing from the loudest voices in the feed. That means you can spot the departments that read but never react, set contribution targets per size band and audience, and report reach with numbers that hold up in front of leadership.
If you want to see what your own reader, liker and contributor shares look like, book a 30-minute call with Jérémy and we will walk you through your network together.
Frequently asked questions
What is the 90-9-1 rule?
It is a rule of thumb published by Jakob Nielsen of the Nielsen Norman Group in 2006: in most online communities, 90% of users are lurkers who read but never contribute, 9% contribute from time to time, and 1% of users account for almost all the content. Nielsen called the underlying phenomenon participation inequality.
Does the 90-9-1 rule still apply to online communities today?
Broadly, yes, but not exactly. A 2014 observational study of four digital health social networks found that fewer than 25% of actors ever posted, and that Superusers produced a weighted average of 74.7% of all content. Content stays heavily concentrated in a small group, but the exact split varies from network to network, so treat 90-9-1 as a heuristic rather than a law.
How many Viva Engage users actually post?
It depends strongly on company size. In Tryane's Viva Engage Benchmark 2026, which analyses more than 3.5 million users, contributors were 10.2% of users in organisations with fewer than 2,500 employees but only 2.4% at 25,000 to 50,000 employees. Readers are the largest group in every size band, peaking at 61% of users at 2,500 to 5,000 employees.
Can you get more employees to post instead of lurking?
You cannot eliminate participation inequality, only shape it. Jakob Nielsen's advice still holds: lower the barrier to contribution, make participation a side effect of tasks people already do, and reward contributors without over-rewarding a tiny elite. On Viva Engage, polls and quick reactions are effective low-effort formats that move likers before they become contributors.
How do I measure employees who read but never post on Viva Engage?
Native reporting already covers it. The Viva Engage Activity report in the Microsoft 365 admin center counts, per user, the messages posted, conversations read and messages liked over the last 7, 30, 90 or 180 days, and you can export the data to a CSV file. The limit is segmentation: user details are anonymised by default and there is no department or country view, which is where dedicated analytics adds a layer.
Is lurking bad for an internal social network?
No, lurking is normal behaviour, not a failure of your communication. Nielsen's warning is different: because posters are a small and unrepresentative group, reading their replies gives you a biased picture of what the whole workforce feels. Measure readership directly, and use low-effort touchpoints to give silent employees a way to signal what interests them.
Sources
Further reading
• How to create relevant Viva Engage communities?
• Average Viva Engage engagement rates by company size
• How to Identify the best time to post on Viva Engage ?
• Viva Engage Post Ideas Are Not Your Problem. The Missing Programme Is
If you want to see what cross-channel reporting looks like against your real data, Tryane runs a 15-minute walkthrough with no slides. Book a slot with Jérémy.