The Guradian’s crowdsourcing experiment

Yesterday at the University of Melbourne’s Future of Journalism conference I used the Guardian’s experiment to use their audience to investigate MPs expenses as an example of crowdsourcing at work.

Thanks to the release of 450,000 pages of receipts in pdf format the overwhelming task of uploding all these documents online was made simple – the whole thing took a week from a software developer, a few days work from others in the department, and a mere 50 pounds to rent temporary servers.

Within the first 80 hours the Guardian audience has investigated 170,000 documents with a staggering participation rate of 56 %. That has slowed but even now 23,000 readers have trawled through 199,000 pages of documents – with the Guardian continually updating the story with new material.

Importantly it’s worth reading this article from the Niemen Journalism Lab on why it worked – as it’s not just a matter of whacking up the documents and letting the audience run riot.

The developer behind the project – Simon Willison – offered four big tips on making it work:

1. Your workers are unpaid, so make it fun.
2. Public attention is fickle, so launch immediately.
3. Speed is mandatory, so use a framework.
4. Participation will come in one big burst, so have servers ready.

He goes into details on all these so make sure you check out the story behind the story – but it’s a stunning example of using simple media technology for citizen journalism and the rich rewards you can reap as a result.