Starting a new type of service site

How to create a vivid community to maintain it?

[lammert]

Since my study at the university in the late eighties, I have been working intermittently on a tough algorithmic problem. It is a type of problem which needs both a fast computer system and a large amount of human provided and quality checked data. Users of the system can be divided in two groups: Users who just pull data from the system, and contributors who help to maintain the quality of both the data in the background, and the output quality of the algorithm.

Users will expect a near real-time performance of the system. The algorithms used to process the data unfortunately don’t perform well on distributed computer architectures and it was therefore until a few years ago that I couldn’t get it running fast enough on available hardware to make such a system usable for a wide public. Times have changed however and with current multi-core GHz processors I am finally at the point that the system is working at an acceptable working speed.

As I mentioned already, the system does not only need a significant computational engine, but also a lot of information provided by humans. In this way you could compare it with systems like DMOZ or Wikipedia, where the personal knowledge of contributors is the only way to achieve the desired quality.

I have now reached a new step in the project where I will buy and install the needed hardware for a production environment (estimated around $10.000) and make it available to the web via a user interface. The problem however will be to interest a large number of people to use it and become an editor or contributor to the system, just like DMOZ and Wikipedia have their contributors and living community. Without going into specific details, the system will have the following specifications:

• The system will be totally working via the internet, with a web-interface for both contributors, and regular users of the system.
  
• The output of the system after the algorithm has done its work is of interest for a large group of people, from different layers in society. It is a question-answer like system, just like for example a search engine (but not being a search engine) and doesn’t need a lot of instructions to learn.
  
• Many people are familiar with this type of system already because it is offered on a small scale by a lot of websites as a subsection of their site. In those cases the information is preformatted and static. The scale on which I would make it available however is unprecedented and needs significant real-time processing. Compare it with static link pages vs. DMOZ, or FAQ pages vs. Wikipedia.
  
• The input and quality control of the system is ideally performed by many different people with different knowledge and skills, just as with DMOZ and Wikipedia.
  
• A contribution of as little of one minute of work of an individual to the system enhances the system’s quality and is therefore a valuable contribution.
  
• I would rather have 400 editors contributing one hour a week, than 10 editors contributing 40 hours a week. Diversity of knowledge and background of people will help ensure the quality of the system, just as with Wikipedia.
  
• The type of information management allows for anonymous contributors, regular editors and category editors with each a different level of access and tools available.
  
• Data entered by contributors will be available in merely unedited form via a GPL type license. The data that has been processed by the algorithm will however only be available in bulk via a proprietary license, or in small chunks via the Q&A interface on the website. Internal discussions about the data will also be private, just as the information in the private DMOZ forums and the edit history attached to URLs in DMOZ.
  
• There will be no ads running on the web interface. I don’t want to scare voluntary contributors away by signs that their contributions are monetized improperly, although it will be known to them that the post-processed form of their contributions may be monetized, which I think is legitimate.
  

Buying and installing the hard- and software configuration seems the easy part, the problem I am facing now is to really get the project flying. This needs a community of contributors to the project and I have no good idea yet how to create and pamper that community to the point that the community and the system can stand on its own.

The system has not enough user contributed content yet to be very interesting for a large group of people. Without user-provided information, few people will use it and consider becoming a contributor. On the other hand, without contributors, the system won’t get the large user contributed information database it needs to become widely known.

How should I proceed further? I do have some money to advertise, but not the money to hire paid contributors for the first phase which according to my estimation needs between 5000 and 10000 man-hours of people with different skills before the system gets interesting for the large crowd.

How did DMOZ and Wikipedia start? How do I build a community of volunteers around a project, which hasn’t much to offer yet, until the input of that community has reached a certain threshold level?

Vandalism seems the main threat to me at this moment. Just as with Wikipedia, malicious editors can cause harm to the quality and consistency of the system. DMOZ has solved this by only allowing selected editors to perform changes in areas that are designated to them. But that is not the way I want to go. How do I reduce vandalism? I can think of a strong moderation policy, sandboxes to start learning the system, limiting changes to the system to be only seen to the editor itself but not to others until the edits are approved, etc.

Any input is appreciated.

[rogerd]

This is the classic community startup problem, lammert - communities without activity don't attract new activity.

Can you try a focus on building out one topic where a modest amount of time would result in a decent-looking knowledge base?

Initially, I'd recommend a high-touch enviroment in which you personally welcome new members, encourage their participation, and recognize their contributions.

Sounds like a fascinating project!

[lammert]

I administrate two forums about other subjects and--although they are not really large--I already have the experience that welcomming new members and encouraging can help to create a stable community.

Those forums were however build around existing topics that attracted visitors already and they had something to discuss about. I now face the problem of having nothing to offer--only an algorithmical engine--until people start to contribute a significant amount. That's why I am interested how sites like wikipedia or dmoz started. Maybe I should take a dive in archive.org to see how these sites looked when nobody knew they existed yet.

I have some knowledge I can add to start up the system, but unfortunately not in an area that would attract a broad instant userbase. It is like I know how to repair a car, but I don't have a drivers license to drive it ;)

[maximillianos]

If you project is as useful and interesting as you say it is, you should not have a problem getting folks to stick around. You find folks who share your passion and interest in the site will want to contribute...

So to get started, go through the regular marketing hoops:

1. Kick start the content with you and friends contributing.
2. Press release
3. Pay-per-click campaigns to drive in some lurkers
4. Keep on top of your stats, see where folks are entering and leaving. Work on keeping the stickiness factor high! Keep conversations going, have social aspects so folks want to stick around and interact.

Obviously it can take years to achieve the success of the communities you mentioned... So just be prepared to spend years of hard work to build it up...

[longen]

I administrate two forums about other subjects

Letting those members know could generate some traffic - from their friends, family, contacts - the network effect.

[rogerd]

Longen's suggestion is a good one. Try to make any promotional posts relevant. E.g., if you want to promote your photography community on your pet community, do so via featuring a photo gallery of pets rather than a simple "check out my new site".

[lammert]

I like Longen's idea. Although my current forums and the new project are on different subjects, there are similarities. Furthermore there is a low barrier between me and the forum members which makes it easier to get them involved.

Very good advice rogerd, the right wordings can make a difference of day and night.