NRS An Overview
The meeting was chaired by Alan Copage and the speakers were Ray Morgan, Roger Beeson and Jean Wong from NRS and Erhard Meier from RSL.
Ray Morgan opened the presentation by saying that the approach they would be taking would be an overview not a whitewash! Roger Beeson continued by outlining the format of the presentation. The issues were to discuss Newspaper sections, the possible solutions and finally an overview of what the NRS has achieved so far.
The presentation then developed into a question and answer session between Roger and Jean (with Jean taking on the role of an agency planner). Jean said that measuring the a.i.r. of parent papers only was no longer acceptable in the current climate when the tabloid/weekend review sections account for 7% of total advertising revenue! Indeed 21% of display advertising revenue is achieved from those section not analysised on the NRS. There are some newspapers sections which earn as much revenue on the own as some independently measured magazines. So why can’t we have readership figures for sections?
Roger Beeson replied by saying that wee cannot just treat sections as bolt on titles. There is a problem of how to measure their readership. Some sections are physically separate from the parent paper and equally some are in the run of the paper. There even some insertions which look like run of paper and some sections have sections within sections! Clear identification of sections is difficult.
On the NRS we currently use masthead prompts, if use the same methodology for Sections readership it may lead to confusion within paper. For example there are 7 separate sections all called “Weekend”. There is also a problem of respondent overload and risk of extending the interview time which is already 35 minutes.
So to summarise, the problems with sections readership research are title confusion within the paper, inability to measure readership reliably, risk of disturbing the main currency and you can only measure recognisable sections which may penalise those less easily identifiable sections.
NRS commisioned BJM to do some qualitative research for them which highlighted the problem of correct identification of sections. In 1988 colour supplements were added to the main survey. The quarter 4 results for that year were suspended due to unreliability. There was an overall 5% drop in readership and in the worst case one title dropped readership by 30%!
Roger then suggested that maybe one approach would be to measure one section only per newspaper. However the recommendation of RSL and Michael Brown (NRS technical consultant) was not favourable. So if readership of sections cannot be measured on the main survey how about running a separate survey?
Roger laid out the brief. What could be done for £250,000? Only the quality and mid market papers should be considered and of those only sections which generated significant display advertising revenue. A solution was to use a recognition based method in a hall test, with face to face interviews. The minimum sample size would be 500 per issue giving a gross sample of 12,500. The major problem with approach was sample size and finding. So where could NRS go from here?
Roger said that the NRS could only do what the industry will accept and posed several questions. Would the industry accept less accuracy, would they pay more, would they consider? The industry must decide and this whole issue is all about compromise. Roger went on to explain that given all the problems, NRS had asked RSL to come up with some possible solutions.
Erhard Meier from RSL, then presented 6 possible solutions with their relevant “pro’s and cons”. The objective in each case was to provide on-going measurement of newspaper sections as part of the NRS. Option 1 was to run a parallel survey. Option 2 was to add to the current NRS sample using page recognition techiques. Option 3 was the RSL initiative and suggested the use of prompt cards and collect data relating to actual issues seen, only validated results would be reported. Option 4 was a re-interview approach which was felt to be costly and would also cause respondent fall off. Option 5 was a reinvestigation of Topic Interest. Option 6 was a survey of attiudes to newspaper sections – however this would not give a useable readership measure.
Roger said that the NRS was not in a position to make a recommedation of any of the options described. The NRS’s role was to provide possibilities not solutions – only the industry could decide on the right solution. He saw three alternative, go for an independent survey with a large sample which would be expensive or accept a change in currency and use the RSL initiative approach or look at further developing the Topic Interest section and lose to some of the product data.
He went on to say that Press research was generally underfunded, particularly by agencies.
Roger was keen not to end on a negative note and finished the presentation by reminding the audience of NRS achievement since 1992. We now have newspaper readership by day of week. Saturday a.i.r. is now published. A model has been agreed to provide consistent data. Early ’95 shoud see the publication of the first 6 months data. Quality of reading has been developed. The measure of primary readership has been modified as will be the source of copy question. Topic Interest has been introduced. There have been major sampling changes since ’92. The main sample has increased by 43% and ABC1’s on the sample have increased. Data is now collected by CAPI leading to faster and more frequent reporting.