How we ran #SaveTheInternet for free

Back in April ’15, the Telecom Regulatory Authority of India (TRAI) released a draft recommendation that threatened the fundamental openness and neutrality of the internet in India.
A few of us got together to try and raise awareness on this issue on Net Neutrality, and it ended up being one the biggest online movements in India in recent times.

We had over 1 million emails sent to TRAI in a span of just two weeks.

I was one of the people behind the movement, and helped build and design the website.

What is surprising, and relatively unknown, is that this entire campaign ran entirely for free. We used only free and open services to plan, coordinate and execute this movement.

This talk is about how we used such tools effectively, and my learnings on how relatively low tech solutions can have far reaching impacts.

Transcript

1. #SaveTheInternet How we ran for free* *not click bait, for real.
2. Karthik Balakrishnan !karthikb351
3. Karthik Balakrishnan !karthikb351 TL;DR - I made the website.
4. The
 Timeline
5. 27 March 2015 TRAI releases their paper
6. 27 March 2015 TRAI releases their paper 1 April 2015 “We’re f*cked” - Nikhil Pahwa
7. 27 March 2015 TRAI releases their paper 1 April 2015 “We’re f*cked” - Nikhil Pahwa 11 April 2015 savetheinternet.in launches
8. 27 March 2015 TRAI releases their paper 1 April 2015 “We’re f*cked” - Nikhil Pahwa 11 April 2015 savetheinternet.in launches 22 April 2015 Rahul Gandhi makes a scene in Parliament
9. April 24 2015
10. April 24 2015 1.05 million emails
    
11. April 24 2015 1.05 million emails
    Wait, what just happened?
12. Just an Email
13. TRAI asked 20 questions
14. TRAI asked 20 questions How hard can 20 answers be?
    
15. TRAI asked 20 questions How hard can 20 answers be?
    Spoilers: Turns out, pretty hard.
16. Low friction, one/two click solution How do you get people to email 20 answers?
 
 

17. Build a form. Simple.
18. Mailgun? Mandrill? hand-rolled mail server?
19. Mailgun? Mandrill? hand-rolled mail server? - SPIF/DKIM
20. Mailgun? Mandrill? hand-rolled mail server? - SPIF/DKIM - Risk of becoming a Change.org campaign
21. Mailgun? Mandrill? hand-rolled mail server? - SPIF/DKIM - Risk of becoming a Change.org campaign - Cannot prove authenticity
22. Mailgun? Mandrill? hand-rolled mail server? - SPIF/DKIM - Risk of becoming a Change.org campaign - Cannot prove authenticity - Money?
23. The Solution
24. The Solution(s)
25. People send the email themselves
26. People send the email themselves Genius, no?
27. Attempt 1 - The mailto: hack
28. Attempt 1 - The mailto: hack - Encode everything in a giant mailto: link
29. Attempt 1 - The mailto: hack - Encode everything in a giant mailto: link - GET request to the compose URL of Gmail/Yahoo/Hotmail
30. Attempt 1 - The mailto: hack - Encode everything in a giant mailto: link - GET request to the compose URL of Gmail/Yahoo/Hotmail - FATAL: URL length limit
31. Attempt 2 - ZeroClipboard.js
32. Attempt 2 - ZeroClipboard.js - Flash hack to copy body to clipboard
33. Attempt 2 - ZeroClipboard.js - Flash hack to copy body to clipboard - Couple with mailto: encoding (minus body)
34. Attempt 2 - ZeroClipboard.js - Flash hack to copy body to clipboard - Couple with mailto: encoding (minus body) - FATAL: Flash on mobile
35. Attempt 3 - Semi-automatic
36. Attempt 3 - Semi-automatic - Step-by-step copy body to clipboard
37. Attempt 3 - Semi-automatic - Step-by-step copy body to clipboard - mailto: link (minus body)
38. Attempt 3 - Semi-automatic - Step-by-step copy body to clipboard - mailto: link (minus body) - SUCCESS: though increased number of clicks
39. Seamless on mobile
40. Seamless on mobile Mobile accounted for 58% of our traffic

41. Seamless on mobile <Insert mobile-is-the-future gyan here> Mobile accounted for 58% of our traffic

42. The Answers
43. 20 detailed, well-reasoned answers
44. 20 detailed, well-reasoned answers They all can’t be the same
    
45. Prevent spamming messages
46. Prevent spamming messages - Lawyers helped draft the answers
47. Prevent spamming messages - Lawyers helped draft the answers - Randomly chose one from a list
48. Prevent spamming messages - Lawyers helped draft the answers - Randomly chose one from a list - Flat doc, easy to edit
49. The amazing pickOne();
50. The amazing pickOne(); Simple and powerful.
    
51. The amazing pickOne(); Simple and powerful.
    Seriously, we used it everywhere.
53. Tech Stack
54. Tech Stack The real reason you came for this talk.
55. The stack is absolutely boring.
56. The stack is absolutely boring. And that’s exactly why it’s noteworthy.
57. Completely static site
58. Completely static site - Styles and code inline, everything else on CDN
59. Completely static site - Styles and code inline, everything else on CDN - Easy to maintain and test
60. Completely static site - Styles and code inline, everything else on CDN - Easy to maintain and test - Grunt slowed us down tremendously - Ditched it after the first couple of hours
61. Completely static site - Styles and code inline, everything else on CDN - Easy to maintain and test - Grunt slowed us down tremendously - Ditched it after the first couple of hours PR still not merged (sorry @SohamKamani)
62. GitHub, no points for guessing
63. GitHub, no points for guessing - Easy to use online editor for answers
64. GitHub, no points for guessing - Easy to use online editor for answers - Lent authenticity - everyone in the campaign was a nobody - me included
65. GitHub, no points for guessing - Easy to use online editor for answers - Lent authenticity - everyone in the campaign was a nobody - me included - Open review
66. GitHub, no points for guessing - Easy to use online editor for answers - Lent authenticity - everyone in the campaign was a nobody - me included - Open review - Anyone could participate, and people did
67. Heroku's free tier single dyno for hosting
68. Heroku's free tier single dyno for hosting - Node.js app serving a static site
69. Heroku's free tier single dyno for hosting - Node.js app serving a static site - Amazing performance, not a single hiccup
70. Heroku's free tier single dyno for hosting - Node.js app serving a static site - Amazing performance, not a single hiccup - Why Heroku?
71. Heroku's free tier single dyno for hosting - Node.js app serving a static site - Amazing performance, not a single hiccup - Why Heroku? - Git deploys
72. Heroku's free tier single dyno for hosting - Node.js app serving a static site - Amazing performance, not a single hiccup - Why Heroku? - Git deploys In hindsight, could have gotten away with GitHub Pages.
73. Issue Tracking
74. Issue Tracking a.k.a. - Why mobile web is broken
 
 

75. Issue Tracking a.k.a. - Why mobile web is broken
 
 
 Seriously broken
76. Google form for issue tracking
77. Google form for issue tracking - Github Issues need you to sign up
78. Google form for issue tracking - Github Issues need you to sign up - We didn’t want to flood it
79. Google form for issue tracking - Github Issues need you to sign up - We didn’t want to flood it - Easy to view issues in a spreadsheet
80. Mission Control
81. Mission Control We just had everyone in a Slack team
 
 

82. Mission Control We just had everyone in a Slack team
 
 
 Nothing more to say other than Slack is amazing.
83. Counting Emails
84. Counting Emails How many were sent via our campaign?

85. Make a Gmail account
86. Make a Gmail account BCC ourselves in every email, easy peasy
 
 

87. Make a Gmail account BCC ourselves in every email, easy peasy
 
 
 What could go wrong?
88. TIL+TIFU - Gmail has an incoming rate limit
89. Solution?
90. Solution? Make more Gmail accounts, distribute mails
 
 

91. Solution? Make more Gmail accounts, distribute mails
 
 
 Wait a second, how many is “more”?
92. Uhm...
93. Uhm... Six accounts.
94. Uhm... Sixty accounts.
95. We thought we’d be lucky with 10k emails
96. We thought we’d be lucky with 10k emails Nothing could have prepared us for 1 million
 
 

97. The “techie” in me would have killed this campaign
98. The “techie” in me would have killed this campaign and that is sort of the “message” of this talk.
 
 

99. Learnings All personal opinions/observations, obviously
100. Speed speed speed
101. Speed speed speed - Viral - You lose people every second
102. Speed speed speed - Viral - You lose people every second - Planning would have resulted in over engineering
103. Speed speed speed - Viral - You lose people every second - Planning would have resulted in over engineering - Reactive approach worked well (somewhat, lost some email :( )
104. The tools are super accessible
105. The tools are super accessible - All this was run entirely on free tools
106. The tools are super accessible - All this was run entirely on free tools - Heroku, Gmail, Github, Slack, Twitter, Facebook, YouTube - all free
107. The tools are super accessible - All this was run entirely on free tools - Heroku, Gmail, Github, Slack, Twitter, Facebook, YouTube - all free - Lack of resources aren’t an excuse anymore
108. People are awesome
109. People are awesome - People from around the country - discussions, contributions, GIFs
110. People are awesome - People from around the country - discussions, contributions, GIFs - Kids from Hyderabad emailed asking if they could use our logo for a poster

111. People are awesome - People from around the country - discussions, contributions, GIFs - Kids from Hyderabad emailed asking if they could use our logo for a poster
 
 - Professor from a university offered to give us college resources for free
112. People are awesome - People from around the country - discussions, contributions, GIFs - Kids from Hyderabad emailed asking if they could use our logo for a poster
 
 - Professor from a university offered to give us college resources for free This warrants its own talk, really.
113. Participatory democracy
114. Large scale impact
115. Large scale impact - We made national news within 48 hours
116. Large scale impact - We made national news within 48 hours - Mark Zuckerberg responded in 72 hours
117. Large scale impact - We made national news within 48 hours - Mark Zuckerberg responded in 72 hours - International news in less than a week
118. Large scale impact - We made national news within 48 hours - Mark Zuckerberg responded in 72 hours - International news in less than a week - Flipkart backed out of Airtel Zero
119. Large scale impact - We made national news within 48 hours - Mark Zuckerberg responded in 72 hours - International news in less than a week - Flipkart backed out of Airtel Zero - Multiple companies exited Internet.org
120. Multi-crore deals thwarted
121. Multi-crore deals thwarted by a 22 year old college student who sometimes misspells Nuetrality 
 
 
 

122. Multi-crore deals thwarted by a 22 year old college student who sometimes misspells Neutrality 
 
 
 

123. Multi-crore deals thwarted by a 22 year old college student who sometimes misspells Neutrality 
 
 
 
 What a time to be alive.
124. Thank You
125. Thank You !karthikb351
126. Thank You !karthikb351 PS: I have got a few stickers left
127. #SaveTheInternet )FMQ
     VT http://bit.do/help-sti !karthikb351