FAT IDEAS

1
How online methods are being used to research
media todayand tomorrow
Richard Thornton, Ciao UK
November 3rd, 2005
2
Why the increasing use of online for media
research?

• Response Rate
• Falling response rates with other methodologies
• Concern for media industry
• Driven by the Media Industry
• NRS, Ipsos, GfK NOP/BBC
• Convenience Methodology
• One-stop project management, international reach,
  global panel coverage
• Self-completion what you want, when you want
• Multi-Media Tools
• Card Sort, Media Evaluators, Brand Trackers,
  Pack/Concept Tests

3
Online Upwards Why the increasing use of
online research?

• Broadband Connection
• - 64 of home internet users (TGI.Net)
• - 50 who dont will be connected in 12 months
• - Expanded multimedia and questionnaire design
• possibilities (Jpegs, Mpegs, Audio)
• Online Activity
• - More activity online, e.g. shopping, banking,
  music
• downloads
• Internet Penetration
• - Internet penetration rates of 70 in some
  countries

4
Online Upwards Internet now mainstream
Source internetworldstats.com
5
Lets ask the Panel What is an online panel?

• Complimentary methodology
• Profiled respondents
• Double-opt in recruitment
• Access to specific target groups
• Quick and cost-effective
• Convenience methodology
• Hard to reach / time pressured respondents

6
Lets ask the Panel What makes a good one?
Detailed profiling to qualify respondents Approp
riate incentive schemes Restrictions on survey
completion Avoid professionalising the
panel Quarantine respondents for sample
extractions Checking procedures to verify
respondents Good panel churn / replenishment
7
Online Picnic No sandwich short here!

Concept Tests
Titles
Media Tests
Product Tests
Ad Tests
Tracking Studies
Omnibus
Opinion Polls
Spots
In home usage tests
Image brand awareness
Blog Sites
Conjoint Analysis
8
Robust, reliable representative samples are
possible!
Even the best panels are biased
Extracted sample representative
500 (50)
185k(58)
500 (50)
130k (42)
Example 1.000 participants, Germany,
representative acc. to gender
Weighting to Census / industry data
9
Ladies Gentlementhe ideal supermodel!
Response rate is affected by a large number of
factors, many relating to respondent ENGAGEMENT
with the survey
Expanded Model
Survey Length
AppropriateIncentives
Panel Involvement
ResponseRate
Respondent Satisfaction
Survey Invitation
SurveyNarrative
Survey Design
10
Online panels offer responsiveness
High response rates can be expected using online
access panels
Response rates taken from last 5 B2B online
project run on Ciaos Proprietary Access Panels
in March 2005
11
OTX Case In Point Communication Research
Objective Test traditional measures of ad
recall, communication, and persuasion Why
Online - More realistic environment
contributes to higher quality outcomes - Provide
turnaround fast enough to be incorporated into
decisions - Create reasonable pricing to
facilitate testing of more
ads/alternatives Vehicle Online magazine test
presenting ads in a distracted environment
12
OTX Case In Point Methodology

• Respondents recruited online to read and react to
  an online version of a current issue of Better
  Homes and Gardens
• Magazine is current issue of typical title in the
  test brands media plan
• Magazine contains 20 total pages including front
  and back cover ( 5 pages editorial / 15
  advertising).
• Although position effects are minor, test
  positions within magazine are kept constant
• Can accommodate spreads, partials and gatefold
  units (in addition to standard 1P ads)
• Virtual magazine allows respondents to read,
  browse, and zoom at their own pace.
• Exposure is within a online survey environment
  that asks about reaction to magazine editorial
  and then asks questions about the advertising.

13
OTX Case In Point Data
14
OTX Case In Point Results

• Unconscious ad exposure allows for use of
  recognition as well as recall to measure
  attention
• Outcome for key communication elements are
  provided for a variety of ads, creating a
  normative database for reasonable cost
• Ability to derive in-depth results for the test
  ads to mirror previous testing methods
• Engagement
• Communication
• Persuasion
• In-Depth Ad Diagnostics

15
Know your limitations

• The Internet is not the answer to all your
  research nightmares!
• Mostly quantitative studies although this is
  changing
• Blogging
• Bulletin boards, qual/quant fusions
• When sample must represent deep census
• Long interviews (25 mins)
• Some demographics under-represented online
• Older age groups
• Ethnic minorities
• Low income / social graders

16
Setting standards for media research
How to create ENGAGING media surveys? its
simple, make sure that you RESPECT your
respondents R Only RESEARCH what you need cut
irrelevant questions E ENGAGE respondents, dont
take them for granted S Take a SATISFACTION
measure to see how you did P PERSONALISE and
humanize it, think about your language E EXCITE
respondents, make it fun and interesting, not
laborious C Dont CHEAT (e.g. telling the
criteria required in the invite email) T Keep up
with the TECHNOLOGY
17
Summary

• There is an ever-increasing usage of online for
  media research fuelled by growth in key Internet
  drivers
• Media industry associations and standards are
  adopting online as a serious methodology for
  migrating existing offline tracking and
  measurement studies
• The challenge will be to bring the best of whats
  being done offline and bring this online without
  losing the reliability of the results