One example:
Another example:
http://www.mcgt.net/index.html
A third example, with more expensive programs besides the kind of program I just mentioned:
http://www.davidsongifted.org/
After edit: To relate what I say here to what was said in the submitted article, an organization that gets all of its members talking to one another freely through one-to-many, many-to-one, and one-to-one channels available to all members gets a LOT of feedback that other organizations don't get, and learns what works and what doesn't work.
There are also meta-charities like GiveWell and the MIT Poverty Action Lab that investigate other charitable attempts.
http://metatalk.metafilter.com/15547/GiveWell-or-Give-em-Hel...
Summaries: http://mssv.net/wiki/index.php/Givewell
They got caught out because they were blatantly astroturfing for themselves and slagging 'competing' charities all over the web. Then it turns out that even by their own generous accounting, more than half of their funds goes directly to their own overhead. On top of that, their 'research' is at best based on getting charities to rigorously assess themselves.
Their rubberstamp board publicly (in the NYT) pretended to fire one of the founders, but less than a year later he was quietly back at his old position.
Do you know of any organisation doing similar but better?
I certainly wouldn't recommend donating to GiveWell, though. The MIT Poverty Action Lab seems like a much better use of the money.
Legal Aid (anywhere) assists asylum-seekers, wrongfully evicted tenants, and indigents needing legal services.
Microfinancing (non-US operations) in India, China, and the Southeast have done so well that the pioneer of microfinancing received a Nobel.