19. TWITTER CASE STUDY
‘GROWTH RELIES ON OPTIMIZATION’
▸ Iterated through many homepage tests
▸ Fixed form validation error with non-
Western account names leading to
+80k new accounts per day
20. PINTEREST CASE STUDY
‘GROWTH HACKING 2015+’
▸ Optimize copy across
customer communications
in 30 languages
▸ Make it easy to run
experiments at scale (30
languages, 15 variants
each)
▸ Have experiments auto-
resolve themselves
https://medium.com/@jwegan_com/how-pinterest-increased-maus-with-one-simple-trick-be424b02429d#.w1n7bpch8
21. Faster Tempo = More Learning
More Learning = More Growth
http://collegesportsblog.dallasnews.com/2014/01/baylor-10th-overall-in-espns-way-too-early-top-25-college-football-rankings.html/
23. ORG STRUCTURE
THE GROWTH TEAM
▸ PM Growth
▸ Analyst
▸ Growth Engineer
▸ Digital marketers
▸ Specialist(s)
Zuckerberg poses with some of his deputies at Antonio’s Nut House in Palo Alto, a popular staff hangout:
Chamath Palihapitya, vice president for growth, mobile and international; Zuckerberg; Mike Schroepfer,
vice president of engineering; Jonathan Heiliger, vice president of technical operations; Chris Cox,
vice president of product.
24. ORG STRUCTURE
HIRE FOR GROWTH
▸ Analytical
▸ Experimental
▸ Fearless
▸ Creative
▸ Determined
Zuckerberg poses with some of his deputies at Antonio’s Nut House in Palo Alto, a popular staff hangout:
Chamath Palihapitya, vice president for growth, mobile and international; Zuckerberg; Mike Schroepfer,
vice president of engineering; Jonathan Heiliger, vice president of technical operations; Chris Cox,
vice president of product.
26. GROWTH PROCESS
IDEAS ARE THE FUEL FOR GROWTH
▸ Full team participation (engineering,
marketing, ops, etc.)
▸ All funnel elements (AARRRR)
▸ Ideas written as hypotheses
27. GROWTH PROCESS
HYPOTHESIS QUALITY DICTATES IMPACT
▸ Data and evidence based
▸ Testable
▸ Defensible
http://mrssirilla.weebly.com/scientific-methods.html
28. GROWTH PROCESS
FOCUS & PRIORITIZATION
▸ Growth lead owns focus
(North Star, OKR, etc.)
▸ Hypotheses are scored and
sized (ICE Score)
▸ Growth team nominates
experiments for testing
Get access to Project @ http://project.growthhackers.com
30. GROWTH PROCESS
PRIORITIZE FOR IMPACT
Base
Conversion
Rate
% Change
vs. Control
Required
Sample Size /
Variant
Test Days
Required
Example A 3% 5% 72,300 72
Example B 3% 10% 18,500 18
Example C 3% 30% 2,250 2
Source: Andy Johns, WMD 2015 talk
35. GROWTH PROCESS
WEEKLY GROWTH MEETING
▸ Growth team participation
▸ One full hour, regimented agenda
▸ Not time for brainstorming
▸ Similar to sprint planning / retrospective
meetings from Agile process
▸ Drives process, accountability, velocity
36. GROWTH PROCESS
GROWTH MEETING AGENDA
▸ KPI Review and Update Focus (15 min)
▸ Review last week’s testing sprint (10 min)
▸ Key lessons learned from completed/
analyzed tests (15 min)
▸ Selected tests for this week (15 min)
▸ Idea backlog review and favorites (5
min)
Get a copy of agenda doc here:
http://bit.ly/growthmtg
Get our testing template here:
http://bit.ly/experimenttemplate
39. SOCIAL LOGIN TEST
Original
Visitors: 3151
Conversions: 605
Conversion Rate: 19.2%
Variation 1 (w/o social sign-in)
Visitors: 3147
Conversions: 750
Conversion Rate: 23.8%
Result: With >99% confidence, the 24% improvement seen with
Variation 1 is statistically significant.
40. FUNNELTEST
Result: With >98% confidence, the 32% improvement seen with 10% variant is statistically
significant. By offering 10% discount we are increasing revenue by 18.9% over control.
10% Variant
Visitors: 4260
Conversions: 303
Conversion Rate: 7.11%
25% Variant
Visitors: 2830
Conversions: 213
Conversion Rate: 7.53%