07 samyagan
- 1. Exposing Non-Traditional Metrics To Run Your Business Sam Yagan & Christian Rudder | OkCupid
- 2. Who is OkCupid? • 4 math majors take on the dating industry • Reinventing business model, user experience, approach • Most innovative dating site (iDate); Top 5 corporate blogs (Mashable)
- 3. Data Drives Daters & Dollars • We’re an underdog; 14 people • Need to compete for consumers with Match & eHarmony – sans 300 employees, sans TV budget, sans Dr. Phil • Need to compete for ad dollars with everyone • Our secret weapon: math
- 4. Sources of Data Demographic data Stated preferences Revealed • Age, race, & • Hundreds of preferences gender millions of • Billions of real questions personal answered interactions
- 5. The Optimal First Message • Question: How can users construct messages that are more likely to attract a reply? • We parsed and analyzed 500,000 messages • We created user messaging guidelines
- 6. !"#$%"&'()(*(+(,&-.( 3,-# Netspeak = u r fail 2,-# 1,-# !"# "# !# $%# &%'# "(%)$# )!*# +%'# 0,-# /,-# .,-# ,-# !"#$%&'&%()'"*$+,'"#-*&'-('!"#$%&'##()'#%.(/-#0/0/1'-%*'+0&-*2' 3*,)("2&4'$+(5*2'#1#0/&-'-%*'#6*"#1*'"*$+,'"#-*'(7'*+,8'
- 7. !"#$%&#'()*+,-"./,0*1,2'3*"(4%&)/, Compliments: ;97% '3"!-4"% *'!/01')12% Body vs. Personality :67% /--+% :97% 867% /()"% ,-.% 897% &"'()*(+% !"#$% 567%
- 8. User Attraction • Question: How does a person’s attractiveness affect his or her success on the site? • This study led us to make some algorithmic adjustments
- 9. !"#$%"&'()$*+,"-"..&/&0$%"&0"..$12-1& Female Attractiveness '!"# &$"# & Male Messaging &!"# %$"# %!"# $"# !"# !"#$%&#'(#)*+"& ,-$%&#'(#)*+"& ()*+,)#-.-/,+0.1#2345637/0.18#+4#6+5)2#79#5:)#*)1# *+,)#;#()*+,)#*)44+<)#2345637/0.1#
- 10. !"#$%&'(")*+$,$--%.%/$0"#$%!$--"12,1% Male Attractiveness '$"# '!"# & Female Messaging &$"# &!"# %$"# %!"# $"# !"# !"#$%& ,"-./,& ,0$%& #'(#)*+"& #'(#)*+"& ()*+#),-)./0+1+22#3425-467/819#)2#-)5+3#6:#5;+#<8(+1# =+()*+#>#()*+#(+22)?+#3425-467/81#
- 11. Wooing Marketers • Dating rarely contextually relevant • Two weapons • Demographics (singles vs. married) • Data
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
- 13. Q: Do you spend more money on food or clothes? Age progression, 18 to 48
- 14. Q: Do you spend more money on food or clothes? Age progression, 18 to 48
- 15. Wrapping Up • For many Fortune 1000 companies looking for social media strategies, data-driven approaches can be easy, effective • We’d love to help you think about your data!
- 16. You’re Welcome! @okcupid | sam@okcupid.com & christian@okcupid.com