Read on Twitter
It’s shipping Friday 
!
Sharing how we set up our product launches that eventually led to 30 million users:
!Sharing how we set up our product launches that eventually led to 30 million users:
1/ None of the following would have happened unless we launched. We launched within a week on Facebook, which was new to third party developers in 2008.
2/ But even with good timing, and phenomenal conditions to grow at that time, work still needed to happen to capture the growth. This is where a good product release process can help. This is what we did:
3/ STEP 1: Set goals for that product release
The goals were always around these 4 areas: acquisition, retention, engagement or monetization
(We focused most on acquisition and retention since our communities were an ad supported business.)
The goals were always around these 4 areas: acquisition, retention, engagement or monetization
(We focused most on acquisition and retention since our communities were an ad supported business.)
4/ STEP 2: Brainstorm and Rank
Monday product meeting to brainstorm & rank features to best achieve our goal.
Rank features by Level of Impact and Level of Effort. Prioritize those that had a high Level of Impact and Low Effort.
Monday product meeting to brainstorm & rank features to best achieve our goal.
Rank features by Level of Impact and Level of Effort. Prioritize those that had a high Level of Impact and Low Effort.
5/ STEP 3: Does this make sense?
This is an important step, once everything was ranked we would take a step back and the PM would facilitate a conversation to check our gut. Don’t let your gut lead these decisions but let it be a backstop.
This is an important step, once everything was ranked we would take a step back and the PM would facilitate a conversation to check our gut. Don’t let your gut lead these decisions but let it be a backstop.
6/ STEP 4: Execute
We’d work with engineering about design and feasibility, make the mocks and by Wednesday we would have mocks done and by Thurs/Friday morning we’d have it ready for release.
We’d work with engineering about design and feasibility, make the mocks and by Wednesday we would have mocks done and by Thurs/Friday morning we’d have it ready for release.
7/ STEP 5: Measure
How would we know we were successful? Analytics feeds into your next set of feature releases. With tools like @segment @optimizely @mixpanel @googleanalytics, no excuse for founders to launch and not know their metrics.
Treat analytics as part of your release
How would we know we were successful? Analytics feeds into your next set of feature releases. With tools like @segment @optimizely @mixpanel @googleanalytics, no excuse for founders to launch and not know their metrics.
Treat analytics as part of your release
8/ Founders can now continuously launch and track these product releases, with tools like @LaunchDarkly. They can design faster with @figma , communicate faster with @NotionHQ , @Jira , etc...
Even a week is too long to launch. Continuous is the standard.
Even a week is too long to launch. Continuous is the standard.
9/ STEP 6: Accountability
Monday morning we’d have an all hands company meeting (<30 EE) where each part of the org would share their progress, presenting a form of accountability for product.
And afterwards we'd repeat the steps with the Monday Product meeting.
Monday morning we’d have an all hands company meeting (<30 EE) where each part of the org would share their progress, presenting a form of accountability for product.
And afterwards we'd repeat the steps with the Monday Product meeting.
10/ LESSONS 
:
LESSON #1: Timing is crucial, but you must also make good use of it
LESSON #2: Analytics is part of the launch
LESSON #3: Setting up a repeatable framework of launches becomes more important than the individual experiments
:LESSON #1: Timing is crucial, but you must also make good use of it
LESSON #2: Analytics is part of the launch
LESSON #3: Setting up a repeatable framework of launches becomes more important than the individual experiments
10/ Every great product starts with a launch.
It doesn’t even have to be a great launch because each launch is an experiment ... just launch.
@justinsmith for teaching us so much cc: @KevinChou
It doesn’t even have to be a great launch because each launch is an experiment ... just launch.
@justinsmith for teaching us so much cc: @KevinChou
