The Future of the National Readership Survey

1
The Futureof theNational Readership Survey
MRG London Conference 3rd November 2005 NRS Ltd
Simon Marquis, Chairman Roger Pratt, MD Steve
Millington, Client Services Michael Brown,
Technical Consultant
2
The Futureof theNational Readership Survey
MRG London Conference 3rd November 2005 NRS
Ltd Simon Marquis, Chairman Roger Pratt, MD Steve
Millington, Client Services Michael Brown,
Technical Consultant
3
Background

• Report from Ocean Consulting Mar 05
• depth interviews with senior management
  IPA/ISBA/NPA/PPA
• Report from Taylor Nelson Sofres
• group discussions planners
• depth interviews buyers
• Series of discussions
• individual NRS Board Directors
• NRS Board meetings

4
Background

• Report from Ocean Consulting Mar 05
• depth interviews with senior management
  IPA/ISBA/NPA/PPA
• Report from Taylor Nelson Sofres
• group discussions planners
• depth interviews buyers
• Series of discussions
• individual NRS Board Directors
• NRS Board meetings

5
Background

• Report from Ocean Consulting Mar 05
• depth interviews with senior management
  IPA/ISBA/NPA/PPA
• Report from Taylor Nelson Sofres
• group discussions planners
• depth interviews buyers
• Series of discussions
• individual publishers and agencies
• each individual NRS Board Director
• NRS Board meetings

6
Background

• Response rate decline
• Titles covered too many? too few?
• Not enough qualitative data
• Too slow to report on market changes
• Whole system constipated, inflexible
• 6. Volatility of data/changes at odds with ABC
• 7. Possibly overstated readerships in some cases

Doing nothing is not an option Ocean Consulting
Report
7
Background

• Response rate decline
• Titles covered too many? too few?
• Not enough qualitative data
• Too slow to report on market changes
• Whole system constipated, inflexible
• 6. Volatility of data/changes at odds with ABC
• 7. Possibly overstated readerships in some cases

Doing nothing is not an option Ocean Consulting
Report
8
Background

• Response rate decline
• Titles covered too many? too few?
• Not enough qualitative data
• Too slow to report on market changes
• Whole system constipated, inflexible
• 6. Volatility of data/changes at odds with ABC
• 7. Possibly overstated readerships in some cases

Doing nothing is not an option Ocean Consulting
Report
9
Background

• Response rate decline
• Titles covered too many? too few?
• Not enough qualitative data
• Too slow to report on market changes
• Whole system constipated, inflexible
• 6. Volatility of data/changes at odds with ABC
• 7. Possibly overstated readerships in some cases

Doing nothing is not an option Ocean Consulting
Report
10
Background

• Response rate decline
• Titles covered too many? too few?
• Not enough qualitative data
• Too slow to report on market changes
• Whole system constipated, inflexible
• 6. Volatility of data/changes at odds with ABC
• 7. Possibly overstated readerships in some cases

Doing nothing is not an option Ocean Consulting
Report
11
Background

• Response rate decline
• Titles covered too many? too few?
• Not enough qualitative data
• Too slow to report on market changes
• Whole system constipated, inflexible
• 6. Volatility of data/changes at odds with ABC
• 7. Possibly overstated readerships in some cases

Doing nothing is not an option Ocean Consulting
Report
12
Background

• Response rate decline
• Titles covered too many? too few?
• Not enough qualitative data
• Too slow to report on market changes
• Whole system constipated, inflexible
• 6. Volatility of data/changes at odds with ABC
• 7. Possibly overstated readerships in some cases

Doing nothing is not an option Ocean Consulting
Report
13
Background

• Response rate decline
• Titles covered too many? too few?
• Not enough qualitative data
• Too slow to report on market changes
• Whole system constipated, inflexible
• 6. Volatility of data/changes at odds with ABC
• 7. Possibly overstated readerships in some cases

Doing nothing is not an option Ocean Consulting
Report
14
The Solution

• The Solution is based on a number of proposals,
  which in due course must be investigated and
  agreed
• These proposals constitute a vision for a future
  NRS the final outcome may (almost certainly
  will?) differ in detail from what is presented
  here

15
The Solution

• The Solution is based on a number of proposals,
  which in due course must be investigated and
  agreed
• These proposals constitute a vision for a future
  NRS the final outcome may (almost certainly
  will?) differ in detail from what is presented
  here

16
The Solution

• Stop using random probability sampling
• Stop using face-to-face interviews

17
The Solution

• Stop using random probability sampling
• Stop using face-to-face interviews

18
The Solution Put the NRS on the Web

• Establish large access pool of web-connected
  individuals (150,000 250,000?) maintained on
  continuous basis
• Each month, select two quota samples, each sample
  being between 2,000 and 3,000 adults
• Split current media list, and measure
  newspapers/supplements on one sample, magazines
  on the other sample
• Use data fusion to produce single NRS database

19
The Solution Put the NRS on the Web

• Establish large access pool of web-connected
  individuals (150,000 250,000?) maintained on
  continuous basis
• Each month, select two quota samples, each sample
  being between 2,000 and 3,000 adults
• Split current media list, and measure
  newspapers/supplements on one sample, magazines
  on the other sample
• Use data fusion to produce single NRS database

20
The Solution Put the NRS on the Web

• Establish large access pool of web-connected
  individuals (150,000 250,000?) maintained on
  continuous basis
• Each month, select two quota samples, each sample
  being between 2,000 and 3,000 adults
• Split current media list, and measure
  newspapers/supplements on one sample, magazines
  on the other sample
• Use data fusion to produce single NRS database

21
The Solution Put the NRS on the Web

• Establish large access pool of web-connected
  individuals (150,000 250,000?) maintained on
  continuous basis
• Each month, select two quota samples, each sample
  being between 2,000 and 3,000 adults
• Split current media list, and measure
  newspapers/supplements on one sample, magazines
  on the other sample
• Use data fusion to produce single NRS database

22
The Solution Put the NRS on the Web

• Establish large access pool of web-connected
  individuals (150,000 250,000?) maintained on
  continuous basis
• Each month, select two quota samples, each sample
  being between 2,000 and 3,000 adults
• Split current media list, and measure
  newspapers/supplements on one sample, magazines
  on the other sample
• Use data fusion to produce single NRS database

23
The Solution Put the NRS on the Web
Initial recruitment of access pool will involve
collection of classification and other
non-readership data short initial
classification interview (15 mins?) and
therefore shorter (monthly) readership interview
(15 mins?)
24
The Solution Put the NRS on the Web
Initial recruitment of access pool will involve
collection of classification and other
non-readership data shorter initial
classification interval
25
The Solution Put the NRS on the Web
Initial recruitment of access pool will involve
collection of classification and other
non-readership data shorter initial
classification interview and therefore shorter
monthly (readership) interview
26
The Solution Put the NRS on the Web

• Establish large access pool of web-connected
  individuals (150,000 250,000?) maintained on
  continuous basis
• Each month, select two quota samples, each sample
  being between 2,000 and 3,000 adults
• Split current media list, and measure
  newspapers/supplements on one sample, magazines
  on the other sample
• Use data fusion to produce single NRS database

27
The Solution Put the NRS on the Web

• Core AIR estimates will continue, but split
  newspaper/magazine samples reduce length of each
  media list, allowing for
• additional titles (supplements, magazines)
• additional qualitative measures

28
The Solution Put the NRS on the Web

• Large access pool will enable short-term
  tactical boosts to standard monthly samples,
  providing quick snapshots of important
  developments in the print medium
• Could also enable publishers and agencies to
  access pool separately, to ask their own private
  questions (at a suitable cost)

29
The Solution Put the NRS on the Web

• Large access pool will enable short-term
  tactical boosts to standard monthly samples,
  providing quick snapshots of important
  developments in the print medium
• Could also enable publishers and agencies to
  access pool separately, to ask their own private
  questions (at a suitable cost)

30
The Solution Put the NRS on the Web

• Introduce smoothing of NRS period-on-period
  estimates
• reduce impact of sampling variation, thus reduce
  major cause of variance with ABC data

31
The Solution Put the NRS on the Web

• Continue to produce measure of AIR, but
• drop recency question (when did you last
  read?) entirely
• develop improved frequency question and frequency
  scale (how often do you read?)
• run new Frequency Calibration Study
• switch AIR estimation from recency to frequency

32
The Solution Put the NRS on the Web

• Continue to produce measure of AIR, but
• drop recency question (when did you last
  read?) entirely
• develop improved frequency question and frequency
  scale (how often do you read?)
• run new Frequency Calibration Study
• switch AIR estimation from recency to frequency

33
The Solution Put the NRS on the Web

• Continue to produce measure of AIR, but
• drop recency question (when did you last
  read?) entirely
• develop improved frequency question and frequency
  scale (how often do you read?)
• run new Frequency Calibration Study
• switch AIR estimation from recency to frequency

34
The Solution Put the NRS on the Web

• Continue to produce measure of AIR, but
• drop recency question (when did you last
  read?) entirely
• develop improved frequency question and frequency
  scale (how often do you read?)
• run new Frequency Calibration Study
• switch AIR estimation from recency to frequency

35
The Solution Put the NRS on the Web

• Continue to produce measure of AIR, but
• drop recency question (when did you last
  read?) entirely
• develop improved frequency question and frequency
  scale (how often do you read?)
• run new Frequency Calibration Study
• switch AIR estimation from recency to frequency

36
Summary of Proposals

• Base the NRS on a Web access pool
• Provide two monthly quota samples of 2,000
  3,000 adults
• Use one sample for newspapers/supplements, the
  other for magazines
• Use data fusion to publish a single database
• Provide for additional titles, additional
  quality questions
• Provide for tactical boosts to monthly samples
• Provide for additional private questions
• Introduce data smoothing
• Switch AIR estimates from recency to frequency

37
Summary of Proposals

• Base the NRS on a Web access pool
• Provide two monthly quota samples of 2,000
  3,000 adults
• Use one sample for newspapers/supplements, the
  other for magazines
• Use data fusion to publish a single database
• Provide for additional titles, additional
  quality questions
• Provide for tactical boosts to monthly samples
• Provide for additional private questions
• Introduce data smoothing
• Switch AIR estimates from recency to frequency

38
Summary of Proposals

• Base the NRS on a Web access pool
• Provide two monthly quota samples of 2,000
  3,000 adults
• Use one sample for newspapers/supplements, the
  other for magazines
• Use data fusion to publish a single database
• Provide for additional titles, additional
  quality questions
• Provide for tactical boosts to monthly samples
• Provide for additional private questions
• Introduce data smoothing
• Switch AIR estimates from recency to frequency

39
Summary of Proposals

• Base the NRS on a Web access pool
• Provide two monthly quota samples of 2,000
  3,000 adults
• Use one sample for newspapers/supplements, the
  other for magazines
• Use data fusion to publish a single database
• Provide for additional titles, additional
  quality questions
• Provide for tactical boosts to monthly samples
• Provide for additional private questions
• Introduce data smoothing
• Switch AIR estimates from recency to frequency

40
Summary of Proposals

• Base the NRS on a Web access pool
• Provide two monthly quota samples of 2,000
  3,000 adults
• Use one sample for newspapers/supplements, the
  other for magazines
• Use data fusion to publish a single database
• Provide for additional titles, additional
  quality questions
• Provide for tactical boosts to monthly samples
• Provide for additional private questions
• Introduce data smoothing
• Switch AIR estimates from recency to frequency

41
Summary of Proposals

• Base the NRS on a Web access pool
• Provide two monthly quota samples of 2,000
  3,000 adults
• Use one sample for newspapers/supplements, the
  other for magazines
• Use data fusion to publish a single database
• Provide for additional titles, additional
  quality questions
• Provide for tactical boosts to monthly samples
• Provide for additional private questions
• Introduce data smoothing
• Switch AIR estimates from recency to frequency

42
Summary of Proposals

• Base the NRS on a Web access pool
• Provide two monthly quota samples of 2,000
  3,000 adults
• Use one sample for newspapers/supplements, the
  other for magazines
• Use data fusion to publish a single database
• Provide for additional titles, additional
  quality questions
• Provide for tactical boosts to monthly samples
• Provide for additional private questions
• Introduce data smoothing
• Switch AIR estimates from recency to frequency

43
Summary of Proposals

• Base the NRS on a Web access pool
• Provide two monthly quota samples of 2,000
  3,000 adults
• Use one sample for newspapers/supplements, the
  other for magazines
• Use data fusion to publish a single database
• Provide for additional titles, additional
  quality questions
• Provide for tactical boosts to monthly samples
• Provide for additional private questions
• Introduce data smoothing
• Switch AIR estimates from recency to frequency

44
Summary of Proposals

• Base the NRS on a Web access pool
• Provide two monthly quota samples of 2,000
  3,000 adults
• Use one sample for newspapers/supplements, the
  other for magazines
• Use data fusion to publish a single database
• Provide for additional titles, additional
  quality questions
• Provide for tactical boosts to monthly samples
• Provide for additional private questions
• Introduce data smoothing
• Switch AIR estimates from recency to frequency

45
The Futureof theNational Readership Survey
MRG London Conference 3rd November 2005 NRS
Ltd Simon Marquis, Chairman Roger Pratt, MD Steve
Millington, Client Services Michael Brown,
Technical Consultant
46
(No Transcript)
47
NRS Overall Response Rates
48
NRS Overall Response Rates
49
NRS Overall Response Rates
50
NRS Overall Response Rates
49
19
51
(No Transcript)
52
(No Transcript)