Why digital analytics?

WHY  DIGITAL  ANALYTICS?

Raymond  Chau  

Nov  2013

�1

About  me

• Building  web  site  for  10  years  

• Working  in  isobar,  a  full  serviced  digital  agency  around  the  globe  

• My  blog:  http://www.whymeasurethat.com

�2

WHAT  IS  DIGITAL  ANALYTICS?

�3

What  is  digital  analytics?

“Collect  and  use  data  to  make  things  better”

�4

How  do  we  collect  data?

�5

PUBLIC  DataPRIVATE  Data

Data

tracking  (e.g.  Google  Analytics)

survey  (e.g.  iPerception)

logs  (e.g.  server  logs)

paid  (e.g.  Hitwise,  Compete)

free  (e.g.  Google  Trends,  forums)

Example:  Google  Analytics

�6

Free  tool  offered  by  Google  to  monitor  user  activities  on  their  web  sites  or  mobile  apps

Example:  Facebook  Insight

�7

Allows  developers  and  Page  owners  to  see  how  their  app  /  page  perform

Example:  Google  Trends

�8

Public  available  data  that  allows  user  to  see  how  a  particular  search-­‐term  performs  over  a  time  period  on  different  regions

Example:  TripAdvisor

�9

Forum  is  one  of  the  best  (and  mostly  free)  source  of  qualitative  data  from  your  customers!  See  how  honest  are  they!!

WHAT  CAN  DIGITAL  ANALYTICS  DO?

�10

Improve  advertising  spending  effectiveness

�11

ROI  can  be  calculated  and  see  which  channel  is  more  effective

Fix  the  entrance  of  the  site

�12

75%  of  visitors  immediately  leave  the  site  after  seeing  the  page.  What’s  wrong?

Optimize  check  out  process

�13

Only  36%  people  checkout  after  adding  things  to  cart,  what’s  wrong?

Where  do  people  go  if  they  quit  from  billing  page?  Are  they  looking  for  shipping  details?

Experimenting  (A/B  Testing)

�14

Personalisation  

• Movie:  Minority  Report  (2002)

�15

Personalisation

e.g.  email  marketing  using  user  behaviour  data  

�16

Personalisation

e.g.  recommendations  based  on  purchase  history  data

�17

Personalisation

e.g.  advertisement  based  on  facial  expression  datahttp://www.cenique.com/cenique/advertising_Platform.php

�18

Recommendations

Provide  recommendations  using  crowd-­‐sourced  data  (e.g.  Netflix,  Amazon)

�19

Re-­‐marketing  /  re-­‐targeting

�20

1.  visitor  enters  your  site

2.  visitor  doesn’t  buy  anything  and  leaves,  but  you  know  he  has  seen  a  particular  product  and  put  it  into  the  cart

3.  visitor  goes  to  a  popular  site,  promotional  ad  targeted  to  the  product  the  visitor  just  read  is  displayed

4.  visitor  goes  back  to  your  site  because  of  the  promo,  and  may  possibly  purchase

TRADITIONAL  MARKETING  V.S.  DIGITAL  MARKETING

�21

Measurability

�22

Traditional  !• Very  difficult  to  measure  • Can  be  very  expensive  to  measure  • Cannot  measure  how  viewers  engage  

with  the  content

Digital  !• Very  easy  to  measure    • Almost  free  to  measure  • Can  measure  how  viewers  engage  or  even  

interact  with  the  content

Message  Delivery

�23

Traditional  !• Message  deliver  to  as  many  audiences  

as  possible  • Have  no  knowledge  on  audiences  • Less  likely  to  respond

Digital  !• Message  deliver  only  to  your  target  audiences  • Know  who  the  audiences  exactly  are  • More  likely  to  respond  as  the  message  is  

relevant  

Example:  Non-­‐target  ad  v.s.  target  ad

�24

Target  Ad  -­‐  Facebook  !• Can  target  visitors  in  finest  level    • For  example,  you  can  target  based  on  

country,  city,  gender,  age  range,  interests,  etc

Non-­‐target  Ad  -­‐  TV    !• Can’t  target  audiences  that  are  really  

relevant  to  your  product  • For  example,  you  can’t  target  those  

who  really  have  acne  problems  

Speed

�25

Traditional  !• Data  takes  time  to  collect  and  analyse  • Cannot  respond  to  audience  

immediately    • Marketers  spend  longer  time  +  more  

money  to  get  insights  and  adapt

Digital  !• Data  are  collected  or  even  processed  at  

real  time      • Can  respond  and  give  tailored  response  

to  audience  real  time  • Allows  marketers  to  fail  cheap  and  fast

LATEST  TOPIC  ON  DIGITAL  ANALYTICS

�26

BIG  DATA!!!

�27

Abacus(old  method)

Calculator  (new  method)

Excel(old  method)

Big  Data(new  method)

Calculation

Data  Storage  &  Analysis

We’re  generating  more  and  more  data

• Example:  facebook  

• 699  million  people  login  Facebook  daily  

• 300  million  photos  are  uploaded  daily  

• Every  60  seconds  

• 510  comments  are  posted  

• 293,000  statuses  are  updated  

• 136,000  photos  are  uploaded

�28

Example:  Obama  2012  election  campaign

• Collect  1  database  that  factors  80  pieces  of  information  about  voters:   (e.g.  age,  race,  sex  and  even  voting  history)  

• Target  right  audience:  focus  on  voters  who  might  change  their  mind  based  on  voters’  ‘persuasive  score’  

• Personalisation:  deliver  policy  messages  or  fund  raising  event  tailored  for  individual  voters

�29

Example:  Disneyland

�30

1.  Opt-­‐in  for  a  wireless  tracking  bracelet  ‘MagicBand’  which  links  to  the  system  My  Magic+

3.  Mickey  Mouse  now  knows  a  lot  about  YOU!!!

2.  Visitors’  real  time  location,  riding  patterns  &  purchases  are  all  recorded

SUMMARY

�31

Data:  to  use  or  not  to  use?

TARGET  ad  reveals  teen’s  pregnancy  http://bit.ly/1b3u5k5  

�32

Want  to  know  more?

�33

ME

twitter  @raymondchau

blog  www.whymeasurethat.com  facebook  fan  page

WhyMeasureThat

Google  +  page  +RaymondChau

LinkedIn  linkedin.com/in/raymondchau

THANK  YOU!

�34