Laserlike

User-generated content and the producer-consumer and producer-producer models.

August 3, 2008 · 3 Comments

If you are building a consumer web site with a user-generated content (UGC) component, chances are that understanding whether you are building a producer-consumer or producer-producer product will make a big difference in the way that you think about your business.

The producer-consumer model.

Traditional media involves a very small group of professionals who produce, edit, and deliver content to consumers.   User generated content came along and changed all of that, right?  YouTube has over 250MM users and Flickr, Digg, and Wikipedia are in the tens of millions.  Not so fast.  People often make the absurd assumption that nearly 100% of users are producers.  That’s not even close to the truth — user-generated content sites typically see somewhere between 0.25% and 5% of users contribute content in a given month.  And that’s not contribute as in contribute quality content.  That’s upload a video, photo, or article.  An even smaller percentage of users actual drive the content that matters to consumers.

 While some sites have managed to keep the cost of participation extremely low and have managed higher levels, I have never heard of a mature site with active participation above 15%.  In fact, I have only heard of one at that level with most closer to <3%.  There are exceptions, but before you name them read the producer-producer section of this post.

The big difference between a traditional media and UGC media producer-consumer model is that you have a much larger pool of contributors and they are often contributing for nothing.  So the user-generated content sites are basically outsourcing the production of content to a very small percentage of users.  This not only lowers costs and exposes undiscovered talent, it offers a diversity of input not previously available to traditional models.  While traditional media companies are still creating a large percentage of quality programming, UGC content captures events everywhere around the clock.

Even the ultimate example of mass-generated content, Wikipedia, is powered by a very small percentage of users – 4,000 editors represent 64,567,607 total edits, with an average of 16,141 edits per editor. This accounts for 32.8% of the 196,705,582 total edits made to the English Wikipedia [according to an article on Wikipedia, of course].

The producer-producer model.

Communications products like email and social networks encourage near ubiquitous participation.  If you count an email as user-generated content, then email is an exception to the argument that a very small percentage of people contribute to UGC.  But the only beneficiaries of an email are the recipients.  With email, the producers and consumers are two sides of the same coin.

While email is unquestionably the largest photo sharing service on the planet (in terms of absolute and relative shares of users sending photos), only the recipients benefit from an “upload.”  Many photo sharing sites provide granular access controls in order to encourage a larger percentage of users to upload content — but you usually end up with either a site that has few people uploading photos with many consuming or behavior that looks much more like email.  Perhaps that is why photo sites with more granular access controls have focused on value-added services like high-end printing?  

The question is where do you put social networks like Facebook?  How about Twitter?

Can you have your cake and eat it, too?

Social networks with certain characteristics manage to have their cake and eat it too — they get very high participation in terms of user content (not just in email-like communications, but in activities where users were previously non-participants).  I believe this is the case for several reasons:

+ Access controls provide security.  Permissions are set through implicit rules (e.g., friends get to see everything by default except for private messages)  Consumers don’t need to explicitly set complex permissions (they just don’t do that) or by implicitly doing so by sending an email or instant message.

+ Participation is easy.  The social graph is an ontology around which participation is easy.  Sharing photos, events, and comments all are linked to at least one person and often numerous people.  You register once and have your email, photo sharing, chat, events, and just about everything with application platforms.

Note that MySpace is likely closer to a producer-consumer model while Facebook is closer to the producer-producer model (if not today, that’s what will happen).

Which model is better?

As I’m sure you already know, it depends.  Nearly all startups I meet assume that they are going to have very high participation.  And most build the types of products which aren’t likely to get more than a few percent of people participating.  

I’ve heard that YouTube has 0.5% of users upload a video in a given month — I would say what they have built is pretty incredible.  Same goes with Yelp, Digg, and Flickr.  If what you are building requires hard work (e.g., putting together a video) or expense (owning a video camera), then it’s a bad idea to focus on large participation.  Instead, focus on finding a small group of loyal and active participants.  Talk to them and get them what they need to make your product serve your consumers — but remember, they aren’t your consumers.  They are closer to employees in terms of their impact on your business.  So treat them very well, but make sure that the product of their work serves your ultimate consumer.

If you are building something that is some type of communications product, you obviously need to target a large percentage of users — if not 100%.  If you are doing something with communications, think about how you can have your cake and eat it too.

Oh, and on Twitter… it may just be a good example of an exception of the producer-consumer model with a large percentage of producers.  What do you think?

Categories: ideas
Tagged: , , , , , , ,

3 responses so far ↓

  • Ben Smith // August 3, 2008 at 11:23 pm | Reply

    I think this may add another rule of thumb metric on social networks. 1% super active, 19% active, 80 percent barely active….and .05% massive contributers to broadly accessed content.

    While the number on our deal are a bit different, we may be just be giving too much credit to the 1-5% and not looking closely enough at the subset of .5% who seem to live for creating daily content. It would also be interesting to see how this applies to things like SixApart and Blogger.

  • Sean // August 4, 2008 at 12:14 am | Reply

    Because it is so easy to be a producer on Twitter (i.e., no more than 160 text characters and no uploading) it makes conceptual sense that it would have a higher producer percentage rate than most other UGC sites. That being said, after a little digging through my own network I found that it was actually lower than what I expected.

    To test the hypothesis I did a quick analysis of the last 200 tweets in my network to understand the producer distribution curve. I found:
    - 49% of tweets were generated by only 10% of contacts
    - 73% of tweets were generated by only 20% of contacts
    - 85% of tweets were generated by only 30% of contacts
    - and 43% of contacts contributed nothing at all

    What does this prove? Well, it is probably too small a sample size to make any definitive judgments about the Twitter producer-consumer ratio overall (after all this is just one’s person’s experience with the last 200 tweets). However, it seems pretty safe to conclude that I’m a geek with way too much free time on my hands on a Sunday afternoon ;-)

  • Estrafalarius — En Twitter también la mayoría es pasiva // June 10, 2009 at 7:20 am | Reply

    [...] generado poor unos cuantos. Por ejemplo, en Wikipedia, un 15% de los usuarios son responsables del 90% de los artículos, pero me sorprendió saber que de acuerdo a un estudio hecho por Harvard Business Review, tan solo [...]

Leave a Comment