Content roles and the Del.icio.us Lesson May 10, 2006
Posted by Steve in : enterprise, tagging , trackbackI’ll soon toss out a lot of my own ideas about how the use of social bookmarking — any or all of the features of del.icio.us, plus more? — could benefit an enterprise. But as I mentioned in my previous post about adoption, creating a seemingly beneficial product doesn’t guarantee that it will actually be used…
A lot of people point to del.icio.us as a great example of how to quickly jump-start adoption, by leveraging the work of early users of the site to make the app increasingly attractive to new users. At first, people were using the system to store their personal bookmarks, then tagging them to organize them into their own taxonomies. Once enough people started tagging, a kind of collective intelligence emerged from the “tag soup”, automatically creating communities around certain tag keywords, which in turn attracted more users (or so it’s said - I’ll comment in a later post about this.)
This is the kind of thing that can only work if an app is really “sticky” even for the first users of the system. And it’s true, you can benefit from use of a social bookmarking system without needing anyone else to participate, but each added participant makes the whole system more valuable. This recognition that personal value precedes network value is now commonly being referred to as the “Del.icio.us Lesson.”
So if you’re hoping for a similarly viral spread of your application, it’s a no-brainer to check any new application for immediate personal benefit to its users. But as obvious as it may be in hindsight, it’s very easy to misapply this lesson during design. After all, most products claim “individual benefit” to some degree, are they all going to take off so quickly? Before we can count on the same kind of feedback cycle working, we at least need to know: who are the individual users, and how will they benefit?
We’ve been tracking participation metrics in our product for over six years. Our first version was a web application with millions of monthly visitors. The more recent enterprise version of our product has been deployed to thousands of employees at multiple F500 companies. Interestingly, both versions showed very similar relative levels of participation: what we found was that for every person who visits, about one in ten sticks around to read more; about one in ten of those signs up; about one in ten of those starts contributing content… and so on. Out of the far end of this diminishing pipeline emerges a community of very active participants who continually add new content throughout the application’s entire taxonomy, which in turn validates the application as a useful source of information for new visitors.
Perhaps that’s why this interesting post by Bradley Horowitz of Yahoo was no surprise to us! His view of Yahoo community content properties (including del.icio.us) is that their potential for success lies in their ability to leverage the work of a small percentage of users (the “creators”) to provide benefit to the rest (the “consumers”). He even claims approximately the same relative participation proportions that we observed, that one in ten users take the next step towards becoming creators. Note that this is not exactly the same as the del.icio.us lesson, but it can be assumed that the users have to first find personal value before they make the effort to advance through the pipeline.
A similar distribution appears in Ross Mayfield’s post about the power law of participation. He presents a great list of the roles in a typical collaboration app: Reader, Commenter, Subscriber, Writer, Moderator, etc. versus the exponentially increasing amount of “engagement” required. I’m not sure every one of them is positioned properly on the chart (and I disagree that this is related to Chris Anderson’s Long Tail - not every curve has a long tail!), but the article is one of the best I’ve read about this phenomenon.
This is a really useful way to think about the users of most social applications: it’s really about scaled levels of content production. How do you encourage viral effects in social sites? Obviously, benefiting users at the more common end of the distribution would directly help the most people! But if your application depends on new user-generated content then it’s important to tend to the publishers, thus indirectly benefiting more users. Or better yet, reduce the effort required to for users to move up the curve.
Does enterprise bookmarking and tagging depend on user-generated content? I think many proponents of del.icio.us and its Lesson would claim that personal bookmarking and tagging is its immediate individual benefit, and that since most of its content is links to external resources, del.icio.us can afford to bypass the content-production end of the spectrum. In my next couple of posts, as I talk about specific features and benefits, I’m going to try to show that none of that is true…
related: del.icio.us lesson
related: collaboration roles
Tags:enterprise tagging
Comments»
Steve, I will be curious to see your next posts on this. Good stuff. Thanks for commenting on my post: Web 2.0 Adoption in the Enterprise - It’s Personal
I’ll be curious to see how well we line up.
Tony
One more question… it is interesting to see the 100:10:1 ratio listed in the article you cite (Yahoo participation). In the world of corporate learning solutions (or really most applications), if you told your customer about that ratio, they would be concerned about the value proposition (given that most people are not participating … and even those that are participate in a less active way). “What about 90% of non-users?”
Do you find that this ratio holds pretty steady in other applications? What’s the answer to when your potential customers ask about the other 90%?
Great question - and yes, we’ve had to deal with this exact issue when selling our product… Any deployment we do has desired usage metrics to meet, so we bring this stuff up proactively to set correct expectations.
But I didn’t mean to imply that 90% of the target audience are “non-users” getting no value from the app! That first “10:1″ ratio of non-users/users can be improved by a lot of things:
- something like the “Perceived Usefulness” concept you blogged about
- the maturity of the site, ie. amount of existing good content
- number of times potential users are reminded of the app
- mandated visits, ie. strong management by app sponsors
- good integration into daily apps
Our enterprise app sees much higher adoption within audiences because we do all of these things. Note that the example in the Yahoo article starts at the “consumer” (ie. already a user) and I guess that’s the real parallel to our observed stats.
Both our internet product and current enterprise product, initially focused on knowledge-sharing via Q&A, had very discrete steps to change from a “reader” to an “asker” to an “answerer”/”expert”. That’s the strongest comparison to the yahoo pyramid.
I think the “10:1″ ratio (which may just be a coincidence) mainly applies where users have to do real work to undertake the next step of engagement, to use the term from Ross’s article. So 10:1’s can be reduced by:
I guess the goal of any app should be to tackle perceived & unmet needs as well as e-mail has fulfilled our general communication needs?
- negating a step, for instance “signing up for an account” becomes a non-issue with enterprise applications, where we always implement single sign-on
- introducing an easier intermediate step
- realizing that your audience is trying to use your app differently than you intended, and adding features to support that behavior
- etc.
But you’re still up against self-selection of roles; at some point you can’t force someone to become a “Writer” if he’s just not interested (or can you? a topic of a later post…)
I don’t have detailed stats from other apps, but clearly e-mail pretty much reaches 100% audience participation, so it’s not some sort of a participation “Law”
Anyway, in the end I think it takes a combination of Ross’s engagement chart and bradley’s magnified-effect chart to really document the value being produced by the community… For me the takeaway wasn’t so much that there are users missing out on value, but that a small number of highly engaged users can provide enough value for a large number of more passive users. This is the same thing we show customers - you might not see as much new content as you expected but take a look at the viewing/searching stats (or better yet, poll your users) and you’ll be surprised!