Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Open Source as a Business (PyCon SG 2014)

Open Source as a Business (PyCon SG 2014)

51567a4f786cd8a2c41c513b592de9f9?s=128

David Cramer

June 21, 2014
Tweet

Transcript

  1. OPEN SOURCE David Cramer twitter.com/zeeg AS A BUSINESS

  2. I work at Dropbox

  3. but this is a story about Sentry

  4. None
  5. None
  6. It started with a "How do I.."

  7. django-db-log (2008)

  8. (so ugly that we have no screenshots)

  9. Basically awful, yet DISQUS found value in it

  10. django-sentry (2010)

  11. None
  12. Still semi-awful, but progress!

  13. Sentry (2011)

  14. None
  15. Finally convinced someone better at design to help

  16. Sentry (Today)

  17. None
  18. Maintained by the Community

  19. None
  20. The company maintains clients in PHP, Python, Node, JavaScript, Java,

    and Ruby
  21. Realistically we only write Python and JavaScript

  22. An unfortunate truth

  23. None
  24. All is not lost!

  25. A large ecosystem of developers Raven.NET chef-sentry-handler heka-py-raven logging (R)

    metlog-raven nagios-sentry pyramid_sentry raven-asc3 raven-cfml raven-cpp raven-csharp raven-erlang raven-go raven-grails raven-java raven-js raven-node raven-objc raven-osx raven-php raven-python raven-ruby raven-sh raven-ssas sentry-assign sentry-bitbucket sentry-campfire sentry-facebook sentry-github sentry-groveio sentry-hipchat sentry-irc sentry-irccat sentry-jira sentry-jsonmailprocessor sentry-notifico sentry-notifry sentry-pivotal sentry-plugin-ipaddresses sentry-sprintly sentry-sprunge sentry-trello sentry-youtrack symfony-amg-sentry-plugin
  26. The value of open source is not in others maintaining

    your code
  27. The community builds things we cannot or will not build

    ourselves
  28. Companies get value in recruiting efforts and visibility in the

    technology world
  29. It should take minimal effort to convince a company of

    open source
  30. Most of my work at Dropbox is completely open source

    because its only beneficial
  31. On To Business

  32. Why start a company?

  33. "You should create an AddOn out of Sentry" - @craigkerstiens

    (Heroku)
  34. "Beer money? That can't be that hard!" - Overconfident me

  35. Three months later I spent Christmas building @getsentry on Heroku

  36. While waiting for Heroku's AddOn validation we decided we could

    collect money using Stripe
  37. Two days later we finally had our first paying customer

    (Feb 28, 2012)
  38. Shout out to @mattrobenolt

  39. (who also wrote raven-js and raven-node)

  40. (and became an easy hire for Disqus)

  41. Our Guiding Principals

  42. #1: Nothing is Free

  43. We must create a sustainable hosted platform, but always remember

    people can host it themselves
  44. #2: Don't Over Charge

  45. We bill based on what costs us money There is

    no per-seat, or per-project pricing
  46. #3: Open Source First

  47. We will not fork Sentry and the only private code

    is our subscription management and billing
  48. #4: Our Ideas are Best

  49. Listen to feedback, but never compromise the platform by adding

    features just because they're requested
  50. "Lean"

  51. Early on our entire mission was simply "Don't spend any

    money"
  52. If you continually take a loss it's hard to prove

    that it's worth driving forward
  53. Heroku helped us get launched by covering our bill for

    the first three months
  54. SoftLayer put us into their incubator program giving us $1,000

    in credit per month
  55. Most importantly we were charging from day one

  56. Open Source is Hard

  57. What your community wants and what a business needs are

    usually different
  58. We need to manage subscription quotas but the self-hosted version

    probably doesn't care
  59. We end up with a very large amount of extension

    points so @getsentry can hook public APIs
  60. At times we just straight up add tooling to Sentry

    assuming no one will ever use it
  61. Catering to Customers

  62. We try to build a product that we love which

    we believe creates a product our users love
  63. It doesn't matter whether you're a paying customer or you

    self-host — users are users
  64. Our belief is that the care we take with our

    product leads to a successful viral and organic growth
  65. Which means we get to build an awesome product that

    anyone can use without restrictions
  66. We're Not a Real Company

  67. We built Sentry at DISQUS entirely because we had problems

    we wanted to solve
  68. It's hard to think of it as a business because

    it feels like we're still just hacking on open source
  69. The entire time my co-founder and myself have been full-time

    employees at other companies
  70. We spend lots of weekends and evenings "working" on @getsentry

  71. That time spent has made some great things possible both

    for DISQUS and contributors
  72. We get to blow the money on fun things, like

    sponsoring events, picking up the bar tab, etc.
  73. "If you do what you love you'll never work a

    day in your life"
  74. Thank You!