One of the nicest things about being involved with mySociety is seeing people in other countries starting similar organisations and building similar websites. After a Skype conversation with another eager hopeful last night, I thought I’d blog a bit about the things I think are most important to know if you’re just starting up. Here goes.
1. Absolutely the most totally essential thing is to be an organisation of amazing, politically minded coders, not an organization employing or contracting good coders*. Their skills are your lifeblood, their ideas your bread and butter, and finding the best civic hackers in your country and building your organization around them is the only path to success. And that means they should be making most of the day to day decisions, not you, you ignorant, arts-degree-clutching clot.
2.Ask the public what they think you should build. Not only will that give you access to more ideas than you have yourselves, it’ll engage people with you. Also, it’ll help you focus on the vital business of building sites that users want, not that YOU want.
3. Keep your cost base low, and put all the money you have into looking after your core staff and being nice to volunteers. Work on building a community of volunteers, even if most of them are really just friends rather than people putting in lots of time. Avoid renting offices, avoid non-essential non-coder staff, get people to donate serving infrastructure and bandwidth. Because building and running democratic websites is a fundamentally new area of human endeavor (not like blogging which at least has an analogy in journalism) there are basically no pre-existing funding streams for the type of work you’re about to do. You will have to create the buzz around yourselves that will lead to people wanting to fund you, and it will probably take years, if you get there at all (mySociety hasn’t quite done this yet, even at 5).
4. Ensure that the core of what you build can struggle on by even if your whole organization collapses. That means being open source, putting energy into sites that are as automated as possible, and making people excited about being volunteers.
5. If you aren’t pissing off at least some people all the time, you’ve probably been captured by the establishment.
6. Take whatever your first website plan is and remove 90% of the features you want. Then build it and launch it and your users will tell you which features they actually wanted instead. Build them and bask in the warm glow of appreciation.
*yes, yes I know I can’t code for toffee, and I’ve got an arts degree, but I’m still a geek, honest. (Proudest moment, working out that a batch of PCI network cards were unreliable because they’d come from the factory flashed with the EPROMs for the wrong hardware, and fixed it.)
Thank you for this brilliant piece !
You are really open-sourcing hard to get experience !!
And it rocks!
Great advice on hoe the models are being evolved, so trust the users to define what is most useful.
I’d add to that that I’d like to see more radical transparency baked in at the beginning.
So, that means yes 1) all software you make is open source, but also 2) all your financial dealings are public (RSS feeds of bank statements, where are they!), 3) voluntarily obey your local Freedom of Information law (unless it is crap, in which case obey a better one).
Put those three in your articles of association – they collectively stop you ever selling out, but they also act as an excellent demonstration to Governments of three principles that they should be following.
At mySociety we do 1) (https://secure.mysociety.org/cvstrac/dir?d=mysociety), we’ve had a go at 2) (http://www.ukcod.org.uk/Finances) but don’t do it as well and as automatically as we could. And 3) feels terrifying; just thinking about it makes you realise exactly how impressive it is that public authorities do it.
Thanks for this great post! All really, really useful. We’re just starting out ourselves (OpenAustralia), so no doubt over the coming months I’ll be checking back to this every so often to remind myself.
Also, what Francis mentioned about total financial transparency is something that we have always wanted to do (not that we have any money yet, but heh, we’re getting there). A couple of things that come to mind – would you disallow anonymous donations completely or would you allow it under the condition that the organisation is ignorant of the donor as well?
Also, the charity’s staff would have their salaries known to all. How are people going to feel about this?
I’ve no experience of how to make it work, so the below is just my guesses. I suspect you can find interesting organisations who have done radical financial openness before.
Obvious suggestions: Cap the amount of individual donations, and show those in summary (total amount, average per donor). Any larger donations would have to be from a named source.
Salaries… These are known to all for people who work for the UK government, in the sense that somebody is on a known pay grade, although you don’t know where within the range they are. You could do something similar – i.e. we spend so many dollars on so many software engineers, in salary ranges X-Y etc.
Or go for radical openness. Consider Mozilla. For key officers there is a lot of salary detail buried in Form 990 http://www.mozilla.org/foundation/documents/mozilla-2006-financial-faq.html, including Mitchel Baker’s $500,000 a year http://thetruthaboutmozilla.wordpress.com/2007/10/22/financials-be-damned/
Perhaps people just won’t mind.
But isn’t point 1 & 2 opposite opinions from a developers perspective ?
pether – no. Because amazing, politically minded coders want to build good ideas that the public want.