Cisco Blogs

Is the 90/9/1 rule still relevant on an Enterprise Collaboration Platform?

December 20, 2011 - 0 Comments

Discussions about online community building frequently turn to a metric called the 90/9/1 rule, first popularized by usability guru Jakob Nielsen. It states that in an online community of user generated content, the breakdown of users looks something like this:

  • 90% are “Lurkers” who read and observe content, but don’t actively contribute.
  • 9% are “Editors”, who modify existing content or leave comments, but never generate content from scratch.
  • 1% are “Contributors” who generate new content/discussions from scratch.

The basic concept of a small group of people generating most of the value in a system is often called “the 80/20 rule”, “the 1% rule”, the “Law of the Vital Few”, “Participation Inequality”, etc. They are mostly are variations of a rule of thumb in statistics known as the Pareto Principle, which states that in many systems, 80% of effects come from 20% of the possible causes.

There are a few important things to keep in mind when applying this rule. First, it is a rule of thumb, not a law of statistics. The ratios can change dramatically by varying the size of the system and the number of causes/effects being considered, and they do not always add up to 100%. Second, it applies to cause and effect analysis, which isn’t inherently good or bad. It may be that 80% of your system failure come from 20% of your bugs, or it may be that 80% of your revenue comes from 20% of your customers. Either way, for the distribution to be useful, it is important to have the causes/effects you are measuring mapped correctly to the things you truly value in the system.

In the older more established forms of internet collaboration (i.e. blogs, forums, wikis) the mapping was simple; content = value. Without new discussion threads, blog posts, and wiki pages, the 90% of lurkers would have nothing to read, search engines would have nothing to index, and the 10% of active/semi active contributors would have no audience and less motivation to contribute. Less content meant less value to both visitors to the site’s stakeholders.

But is this still the case? Many of our daily activities that used to be purely offline are now represented as online social activities that show up in our friend’s activity feeds. Simply walking into a restaurant with the Foursquare app activated on your phone can generate a “check-in” activity, with little to no involvement by the user. Checking off tasks on a collaboratively managed to do list powers the activity feeds in Basecamp, a popular web app for project management. These activities don’t necessarily generate content or discussion, but none the less are used to build the social profile of users, and contribute to the value users get from the system as a whole.

What do we truly value in an Enterpise?

A quick review of the latest consumer research shows that larger sites such as Facebook are already noticing this shift. In Data Informed, Not Data Driven, Adam Mosseri (a product designer at Facebook), notes that as more everyday activities get represented in a social network, the more users can be counted as “active participants”… but that determining the true value of their contributions becomes more difficult. In the case of Facebook, data showed roughly 25% of users logging in daily, commenting, sharing, and posting new content (note it is not 1% or 10%, as the 90/9/1 rule would predict). To grow the ranks of the 25%, they formed an “Engagement Team”, whose first experiment was the “Like” button. The goal of the “Like” button was simple; boost participation overall, as measured by server read/writes. By that metric, the button was a resounding success, boosting traffic by 11% the day it was released, and providing Facebook with a trove of new data about their most active users. But in terms of converting lurkers into engaged users, it did very little.  Mashing the “Like” button was fun an easy, but not something that a lurker would see value in. Birth announcements, wedding photos, meaningful exchanges with friends; those were the things that were most valuable things on the network, that could drive deeper levels of engagement.  From this they concluded that “it is very difficult for a single set of metrics to represent what you truly value”.

When designing an enterprise social networking system this holds especially true; people in companies do all kinds of work that doesn’t involve writing blog posts or updating wikis. One of the primary design goals of an enterprise collaboration system should be to find ways to represent all activities that generate business value on the company’s social graph. Only then can we shift our traditional adoption metrics to start reflecting the new reality of social networking; that having only 1% of your users actively contributing to the enterprise network is actually terrible. Take day to day work actions like fixing bugs, crossing tasks off a project schedule, closing support cases, landing the big sale, or even passively listing expertise in your personal profile; these activities may not generate web content in the traditional sense, but they inarguably generate value for the company. Being able to represent these actions and infer from them which people are contributing the most value to the network will be where the true promise of enterprise social networking is realized. When we can view a co-worker’s activity feed on their profile and immediately understand how they contribute to the company, that is when knowledge management in the enterprise becomes truly powerful.  The more types of work the system can represent, the less relevant the “content = value” baseline becomes.

For this reason, we believe that the 90/9/1 participation rule is no longer a valid means of gauging adoption success in an enterprise collaboration system; it sells the promise of enterprise collaboration short and sets the bar artificially low. As more work-related activities become represented on the social graph, the number of lurkers should drop dramatically, to the point where being a lurker means you aren’t doing any work.

Coming Soon: A new model for measuring adoption success in an enterprise social network

In an effort to keep conversations fresh, Cisco Blogs closes comments after 60 days. Please visit the Cisco Blogs hub page for the latest content.