codeforamerica / cityvoice

A place-based call-in system for gathering and sharing community feedback
MIT License
47 stars 35 forks source link

v1 CityVoice Instance Builder: "Create your own CityVoice instance in 15 minutes for free" #209

Closed daguar closed 9 years ago

daguar commented 10 years ago

Overview

I've got a working alpha to automate CityVoice deploys on Heroku — right now, just the default codebase and content gets deployed — here: https://github.com/daguar/cityvoice-builder-heroku/ http://cityvoice-heroku-builder-dev.herokuapp.com/

Below I've spec'd out what I think a v1 of the "CityVoice Instance Builder" micro-app might look like below.

My core notion here is that for v1 builder, we do not attempt to make something that would be your production deploy — instead, we present a simplified interface leaning heavily on defaults to let someone get a instance that has the key stuff customized quickly. The hypothesis is that if users can get something up quickly, they'll be more likely to go do the (non-trivial) work of a full deploy (e.g., careful survey design, nicely-recorded audio files, etc.)

The goal is more to quickly convey (via the experience of building it) the different pieces they'll need for a CityVoice instance, and also experience what their deployment will be like.

Additionally, my thought is by starting with this less ambitious version, we can learn more quickly where the yaks in a CityVoice-Builder will be (e.g., is HTML5 audio recording terrible?)

In sum, I think this approach lets us sprint to a v1 that adds value today — "do it in 15 minutes" — rather than a partially-built builder that won't be usable until it's fully operational (Death Star, wah-nah-nah).

Proposed v1 spec

Locations

daguar commented 10 years ago

Two additional concrete things to do that might be helpful for this:

migurski commented 10 years ago

I support this approach. I think the difference between "quick" and "real" might be illusory, and if we can get someone up and running in 15min then maybe there is nothing further? Particularly, it'd be great to make it easy to rapidly release a survey, try it out, and then modify and extend or fix it.

jmadans commented 9 years ago

BLUF: With the right client support, City decisionmakers are willing to launch with a "quickstart-ed" cityvoice campaign.

While the jury is still out on on our 5 midflight CityVoice pilots, the core assumption Dave made about the power of "deploying a thing in 15 minutes" has been tentatively validated. The value of letting government make a thing has -in large part - contented them to work within the default constraints of the builder.

In fact, the functionality of the builder has somewhat outperformed Dave's original hypothesis that doing a quick, dirty prototype survey could be the bait to get a city to commit to a production deployment of City Voice. Instead, we have found that augmenting the builder experience with targeted client support (survey design, instance customization, project management) has yielded CityVoice campaigns that government officials are willing to launch to their residents.

migurski commented 9 years ago

For posterity, a link to the builder: https://cityvoice-setup.codeforamerica.org