• Share
  • Email
  • Embed
  • Like
  • Save
  • Private Content
Big Ideas from Big (or Small) Data
 

Big Ideas from Big (or Small) Data

on

  • 6,160 views

Presented at Book Summit Canada, June 2014. How to identify, understand, and efficiently grow your audience by gathering and utilizing consumer data. Tools, techniques, and actionable insights in this ...

Presented at Book Summit Canada, June 2014. How to identify, understand, and efficiently grow your audience by gathering and utilizing consumer data. Tools, techniques, and actionable insights in this presentation, which takes its focus a hypothetical challenge of growing the audience for Nate Silver's book The Signal and the Noise in Canada.

Statistics

Views

Total Views
6,160
Views on SlideShare
5,191
Embed Views
969

Actions

Likes
13
Downloads
55
Comments
0

14 Embeds 969

https://twitter.com 466
http://www.mccarthy-digital.com 441
http://logicalmarketing.tumblr.com 17
https://www.tumblr.com 15
http://moodle.urse.edu.mx 12
http://cgyukna.pbworks.com 8
https://tweetdeck.twitter.com 3
http://mccarthy-digital.tumblr.com 2
https://www.linkedin.com 2
http://pulse.me&_=1405145552747 HTTP 1
http://pulse.me&_=1405145553285 HTTP 1
http://logical-marketing.com 1
http://www.slideee.com 1
http://www.google.com 1
More...

Accessibility

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

    Big Ideas from Big (or Small) Data Big Ideas from Big (or Small) Data Presentation Transcript

    • Big  Ideas  from  Big  (or  Small)  Data   Book  Summit  Canada       Pete  McCarthy   The  Logical  Marketing  Agency  
    • Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada    June  19,  2014   2   Who  am  I  and  why  am  I  here?  
    • Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada    June  19,  2014   3   What  are  we  talking  about  and  why  are  we  talking  about  it   (now)?  
    • We  are  talking  about  big  ideas.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     4   Really,  a  process  which  may  yield  big  ideas.  Discussion  of  data  is  highly  probable.     It  is  a  capital  mistake  to  theorize   before  one  has  data.  Insensibly  one   begins  to  twist  facts  to  suit  theories,   instead  of  theories  to  suit  facts.   –  Sherlock  Holmes,  A  Scandal  in  Bohemia  
    • This  is  a  big  idea!   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     5   94%  accuracy  of  opening  weekend  box  office  up  to  4  weeks  pre-­‐release…   2013  
    • So  was  this  and  seems  to  still  be.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     6   97%  correlation  between  “Twitter  chatter”  and  opening  weekend  box  office.   2010  
    • Especially  when  combined  with  this  work.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     7   Which  adds  (a  little)  more  (seemingly  correct)  data  to  eliminate  bias.   2012  
    • This  might  be  part  of  a  big  idea…   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     8   77%  “predictive.”  Backward-­‐looking.  Reliability  of  data?   2012  
    • 2013   1983   These  were  big  ideas…and  some  still  are…   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     9   Most  big  ideas  build  on  prior  big  ideas  –  successful  or  not.   2010   2010   2002   2000   1994  
    • Why  we  are  here.     June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     10   Because  of  what  Google  (and  others)  do.  Because  we  can  do  similar  things.   ü  What   ü  When   ü  Where   ü  Which   ü  Who   ü  How   ü  Even  a  plausible   why!  
    • Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada    June  19,  2014   11   What  we  talk  about  when  we  talk  about  consumer  data  
    • In  essence,  we  are  talking  about  useful  research.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     12   Some  “types”  of  consumer  research  and  the  methods  used.   Secondary   Industry-­‐specific   Qualitative   Non-­‐transactional   Snapshot  in  time   Bricks  &  Mortar   Unknown  People   Unknown  Person    Primary   “Whole  World”     Quantitative   Transactional   Trended   “Digital/Online”   Known  People   Known  Person             |   |   |   |   |   |   |   |   Types  of  Research/Data   Methods  of  acquiring  research  data     1.  By  surveying  people   2.  By  observing  them    
    • Research  that  yields  data  on  audiences  to  solve  below.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     13   Big  data,  little  data  –generally  pretty  similar  data.  Just  scale  and  use  differ.   Aware  &  Will   Buy.   Aware  &  Will   Not.   Unaware  &   Just  Might!   Unaware  &   Just  Fine.   This  is  the  gold  mine  of  readers.  It  is  the   crossover  hit.  Especially  true  for  niche  and   vertical  publishers.   A  must.  
    • Content  created/consumed  by  consumers.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     14   Mary  Meeker  referred  to  the  “data-­‐creating  consumer”  as  a  top  2014  trend.  
    • Major  social  platforms  total  registered  users.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     15   0   200   400   600   800   1,000   1,200   1,400   2004   2005   2006   2007   2008   2009   2010   2011   2012   2013   Millions   Facebook   Twittter   Google+  (Gmail)   Pinterest   Instagram   Registered  users  as  of  May  2013.  Reported.   Several,  culled  by  Search  Engine  Journal  
    • US  social  network  penetration  by  age  +  mobile.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     16   As  of  May  2013.  Via  survey.   Pew  Research:  Social  Media  Update  2013  via  Search  Engine  Journal  
    • Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada    June  19,  2014   17   Canada-­‐specific  data.     Search  Market  Share   June  2014  opt-­‐in  panel.   June  2014.   Top  Social  Media  Sites  Used  in  Last  Month  Canada  “Digital”  Snapshot  Data   Source:  Experian  Hitwise  Canada   §  86%  internet  penetration   §  76%  mobile  internet  penetration   §  56%  smartphone  penetration   §  77%  of  owners  research  products  on   phone,  27%  buy  on  phone   §  82%  Social  Media  penetration   §  55%  Facebook  penetration   §  <2  hours/day  social  media  use   0%   10%   20%   30%   40%   50%   60%   Pinterest   LinkedIn   Google+   Twitter   Facebook  
    • Canada  and  the  U.S.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     18   Sources:  PWC  Global  Media  Outlook,  Census  Data,  Global  Web  Index   Wave     60   7   0   20   40   60   80   U.S.   Canada   137   17   0   50   100   150   U.S.   Canada   254   30   0   100   200   300   U.S.   Canada   315   35   0   100   200   300   400   U.S.   Canada   Population  (M)     Ratio:  1:9     Internet  Users  (M)  Ratio:  1:8.5     Facebook  Users:  Last  Month  (M)     Ratio:  1:8     Twitter  Users:  Last  Month  (M)    Ratio:  1:8.5     Trade  Book  Sale  Ratios   Range  from  1:15  to  1:10…     No  “apples-­‐to-­‐apples”   data  but  directionally   these  provide  a  sense.   A  sense  of  proportion.  
    • Canadian  book  consumers  and  retail.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     19   2012−2013.  Primarily  via  survey.  (I’ve  focused  on  the  Business  category.)     •  68%  Business  book  buyers  =  male    ! >  50%  awareness  =  online    ! Only  20%  purchase  impulsively.   BookNet  Canada,  “The  Canadian  Book  Consumer  2013”    
    • June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     20   Some  really  useful  places  to  gather  consumer  data.   §  Social  Graph   They  know  consumers.  Online   and  offline.  360-­‐degree  view.     §  Ad  Platform     Open  (APIs,  Tools),  app   development,  Oauth  site  sign  on.   §  Constant  A/B  testing   Fail  fast,  fix.     §  Result:  Happy  Users/Advertisers   Despite  incredible  concerns  over   privacy.  Relevance  trumps  it.   §  Search  (&  lots  else)   Massive  share.  YouTube.     §  Ad  Platform   Targeted  inventory  at  an  all   time  high.   §  Literally  Building  a  Brain   Yes.  All  products  data-­‐driven.   Predictive.   .     §  Open   APIs  and  tools.   Oauth  site  sign  on.   §  Massive  growth   Wild  adoption  and  usage.     §  Ad  Platform   Targeting.   §  Timely   Almost  “now.”  Predictive.     §  Open  (for  now)   Can  get  at  the  data.     Oauth  site  sign  on.  
    • A  sampling  of  useful  tools.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     21   Social  Analytics   §  Simply  Measured   §  SproutSocial   §  Social  Bakers   §  Followerwonk   §  Commmun.it   §  Bit.ly   §  Topsy   §  Social  Mention   §  Facebook  Ad  Interface   §  Facebook  PowerEditor   §  EdgeRank  Checker   §  SimplyMeasured   §  Twitter  Ad  Interface   §  Radian  6/Crimson   Hexagon   §  HootSuite     §  Facebook  Insights   §  LinkedIn  Analytics   §  Instagram  Analytics   §  Etc.   Web/Email   Analytics   Web/SEO   §  Raven   §  Compete   §  Quantcast   §  SEO  Quake   §  SEM  Rush     §  Google  universal  analytics   §  WordTracker   §  WordStream   §  Amazon  comp  authors   §  Librarything  tags/ comps   §  Etc.   §  Google  Analytics   §  Omniture   §  ExactTarget   §  MailChimp   Mostly  not  huge,  costly  a  la    Adobe  or  Salesforce   §  Optimizely   §  Etc.   And  many,  many  more   to  fit  nearly  any  use  case   §  Google  Trends   §  Google  AdWords   §  Moz   §  Soovle  (autocompletes   in  general)   §  Seorch  
    • I  like  how  this  guy  talks  about  research  and  data.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     22   Nate  Silver.  (I  like  others,  also).   …if  the  quantity  of  information  is  increasing  by  2.5   quintillion  bytes  per  day,  the  amount  of  useful  information   almost  certainly  isn't.  Most  of  it  is  just  noise,  and  the  noise   is  increasing  faster  than  the  signal.  There  are  so  many   hypotheses  to  test,  so  many  data  sets  to  mine—but  a   relatively  constant  amount  of  objective  truth.   Photo:  Marius  Bugge   Bayes’ Theorem
    • Foxes  gather  “big  ideas”…quickly.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     23   Photo:  Marius  Bugge   “The  fox  knows  many  little  things,  but  the  hedgehog  knows  one  big  thing.”   Hedgehogs  are  Type  A  personalities  who  believe  in  Big   Ideas—in  governing  principles  about  the  world  that   behave  as  though  they  were  physical  laws  and  undergird   virtually  every  interaction  in  society.     Foxes,  on  the  other  hand,  are  scrappy  creatures  who   believe  in  a  plethora  of  little  ideas  and  in  taking  a   multitude  of  approaches  toward  a  problem.  They  tend  to   be  more  tolerant  of  nuance,  uncertainty  ,  complexity,   and  dissenting  opinion.  If  hedgehogs  are  hunters,  always   looking  out  for  the  big  kill,  then  foxes  are  gatherers.  
    • One  second  on  Bayesian  statistics.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     24   No  test  (I  wouldn’t  pass).  The  governing  principle  is  the  thing.   »  Bayesian  statistics  is  a  subset  of  the  field  of  statistics  in  which  the  evidence  about   the  true  state  of  the  world  is  expressed  in  terms  of  degrees  of  belief  or,  more   specifically,  Bayesian  probabilities.       »  Bayesian  statistics  (if  only  practiced  in  spirit)  sets  one  up  to:     §  Statistical  inferences   §  Statistical  modeling   §  Design  of  experiments   §  Statistical  graphics   §  Be  human  (encouraged)   §  Move  quickly,  get  lots  of  data   §  Admit  bias  but  try  to  verify   §  Change  tack  as  indicated   §  Becoming  “less  wrong”  (testing)   §  Becoming  even  less  “less  wrong,”  over   time   §  Demonstrating/validating   We  verify  or  discover  the  big  ideas,  as  opposed  to  just  having  them.  
    • Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada    June  19,  2014   25   Identifying  and  understanding  audiences  using  data  
    • I  wonder  how  The  Signal  and  the  Noise  is  doing?     June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     26   #1  Bestseller.  In  Statistics  Textbooks….   #989  overall.  Without  being  able  to  see  POS,  I  don’t  know  if  that  signifies…   I  might  throw  a   “Business  BISAC”  at   Amazon.  It’s  not  a   textbook.  
    • Nate  Silver’s  audience.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     27   Wonder  who  they  are.  I  have  guesses  but  that’d  be  bias.  Let’s  look.   720k  is  a  hefty  Twitter  following.  He’s  tweeted  often  and  “on  message.”  Recency.    
    • Where  do  they  live?   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     28   Home  locations  of  unnamed  Silver  Twitter  followers  based  on  a  sample.  Directional.   New  York,  LA,  London.  Is  that  Canada  I  see?  
    • Canada?   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     29   It  is  indeed.  But  those  followers  are  in  Seattle.  Drats!   Why  no  Canadian  followers?  Bug?  Opportunity?  (We  know  Canadians  use  Twitter.)    
    • Google.ca  auto-­‐prompts  me  at  “s.”  That’s  good.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     30   1   1b   Book  results  are  low  and  related.   Amazon  is  first   book  result.  Way   below  the  fold   on  any  device.  
    • How  does  the  book  look  an  Amazon.ca,  Kobo,  Indigo.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     31   There  a  book  audience  but  it  feels  small.   Two  reviews    feels  low…   Good  position.   Seem  like  more  consumer-­‐ aligned  categories   Would  have  expected  him   to  be  prompted  above     Nate  Southard…    
    • What  is  the  search  interest  like?   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     32   Canada  –  Spikes  –  Volume  is  on  Him   Interest  falls  but  stays.  Book  present.   Google  Trends  Canada,  US.   January  2007  –  September  2012     September  2012  –  May  2014     Interest  falls  fast.  No  book.   January  2007  –  September  2012     US  –  Very  Similar    
    • Comparing  raw  search  volume.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     33   Canada  Brand  Search  Volumes   US  Brand  Search  Volume   1,400  reach  in  Facebook  CA  advertising    vs.  62,000  in  US     Ratios  feel  as  if  he  is  punching  below  weight.  
    •  More  data  on  interest  in  Canada  allows  inference…   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     34   Silver  does  not  enjoy  the  interest  here  that  he  does  in  the  states.   3%  is  too  small  number,  given  expected  ratios.   Canada  has  about  the  population  of  California.   Hypothesis:  he  is  under-­‐ indexing  in  CA.   Perhaps  there  is  room   for  sales  growth  –  in  and   using  social.  
    • Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada    June  19,  2014   35   Efficiently  growing  audiences  using  data  
    • Mine  adjacencies.     June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     36   Some  potential  adjacencies  for  Nate  Silver.  
    • One  adjacent  audience:  Moneyball.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     37   Google  Adwords  and  Facebook  confirm  connection  and  show  Canada  reach.   =   =   50,000   196,000  
    • Ride  big  waves.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     38   Google  Trends  Canada.   “538”  is  Silver’s  recently  re-­‐launched  site,  covering  things  from  sports  to  politics.   There  is  Canadian  search  interest  in  538.   He  is  predicting  the  World  Cup  winner  in  real  time.   15M  Tweets  on  World  Cup  in  past  month.     The  World  Cup  is  big  in  Canada  (I  did   verify).  Though  it  is  an  adjacency  that  is   further  away,  Silver  has  tied  himself  to   the  World  Cup  explicitly.     Hypothesis:  It  can  likely  be  capitalized   on  to  get  people  interested  in  him.  
    • Reaching  “look-­‐alikes”   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     39   Some  characteristics  of  his  audience.  
    • Regionality  gleaned  from  search.     June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     40   Are  there  attributes  of  the  US  locales  that  “match”  Canadian  locales?  (DMAs)  
    • Comp  authors:  adjacent  fans  and  look-­‐alikes.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     41   Authors  whom  the  consumer  comps,  as  opposed  to  us.  Preferably  outside  book  spaces.   The  intersecting  folks  are  a  great  source  of  look-­‐alike  attributes.    
    • Comp  authors:  adjacent  fans  and  look-­‐alikes.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     42   We  can  use  the  Venn  to  find  people  to  target  who  look  exactly  like  the  shared  followers.  
    • Thinking  in  terms  of  optimizing  “funnels.”   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     43   Goal:  sell  The  Signal  and  the  Noise  in  Canada.  One  potential  funnel  (to  test).     Segment     §  Male   §  Like  Moneyball   §  And  topics   related  directly   to  Moneyball   Platform     §  Facebook   §  Mobile  stream   Landing     §  Kobo  page   Creative     §  A:  Sports   §  B:  Business   This  is  funnel  A.    There  should  at  minimum  be  a  B,  testing  with  at  least  one  variable  changed.   Measure  costs  to  reach  fans  and  conversion  to  sale  (the  goal  here).   See  who  is  responding,  adjust  (more  hypotheses)  or  “get  out.”  
    • This  may  not  be  a  “big  idea.”   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     44   But  if  it  were  to  be  successful  it  would  be  a  nice  one-­‐off  and  could  lead  to  learning   how  to  develop  a  process  of  outsizing  “American”  authors  in  Canada.   »  One  could  systematically  identify  US  authors  with  works  on  sale  in  CA   §  Look  for  the  delta  in  unit  sales  between  US  and  CA.  IF  greater  than  norm,  examine.   »  Do  the  same  with  authors  with  major  digital  presences  in  US  without  in  CA.   §  See  what  can  be  modeled  in  CA  from  the  US  presence   And  so  on…  
    • Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada    June  19,  2014   45   Suggestions  if  you’d  like  them   (along  with  2  warnings)  
    • Suggestions   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada    June  19,  2014   46   »  Establish  goals  regarding  audience  identification.   §  What  outcome  would  be  ideal.   »  Involve  organization  around  the  approach.   §  Marketing,  sales,  publicity,  IT  need  to  align  to  gain  maximum  value.   §  Affects  everything;  physical  distribution,  ad  creative,  PR  to  metadata,  etc.   »  Recognize  that  it  is  a  process  of  testing  and  learning.   §  Failure  (of  a  reasonable  hypothesis)  is  not  a  bad  thing.   »  Buy,  build,  find,  learn  the  systems  to  support  the  work.   §  Capture  learning  at  all  times.   §  Scale  when  the  value  is  there  (eg.  Big  Ideas  are  coming  and  are  repeatable).   May  prove  useful  if  data-­‐driven,  audience-­‐centric  marketing  is  of  interest.   See  warnings.  
    • Two  warnings   1.  This  is  relatively  technical  work  but  does  not  require  one  to  be  a   “data  scientist.”  Just  unafraid  of  technology,  curious,  and  able  to   employ  the  logic.     2.  The  more  one  does  it,  the  faster  it  goes.  It  is  not  fast  at  first  but  is,   in  the  end,  likely  more  efficient  and  will  yield  big  ideas.   June  19,  2014   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada     47   —Nate  Silver,  The  Signal  and  the  Noise    
    • Thank  you   Big  Ideas  from  Big  (or  Small)  Data    |  Book  Summit  Canada    June  19,  2014   48