‘The Digital Revolution’ was the focus of the conference: not only measuring contact with publisher content in all its forms, but understanding more about the nature of digital reading, and the implications for publisher brands and the advertisers who use them.
Alongside the digital revolution, and partly prompted by it, the evolution of print measurement is taking shape in some countries. A number of surveys are taking steps to move beyond measuring readership of the publication as a whole to provide an estimate of advertising exposure within the publication.
Many ‘readership’ surveys now measure publisher websites alongside print publications. According to Erhard Meier’s opening summary of ‘What’s New around the World’ just under half of the readership surveys around the world include such measures, enabling users to see the overall net reach of the publisher brand and an estimate of multi-platform readership.
However, one could not help but have sympathy for some of the dilemmas in reconciling the different online estimates generated as expressed in Steve Douglas’s paper ‘Online Measurement – Too Many Numbers/Relationships’.
An alternative path is to bring together estimates for print publications and websites from different surveys by means of fusion. There were papers on this subject from the Netherlands, Brazil, Hungary, Great Britain, France and the United States, the latter on the integration of print, Internet and TV estimates.
How does online reading differ from reading in print?
Going beyond measurement, there was a real push to understand more about the nature of online reading, and its implications for advertising effectiveness, with many papers on this theme.
Judy Vogel and Bob Shullman spoke on the topical subject of ‘Viability of Charging for Online Content’. They believe that the opportunities to charge will come in particular from mobile platforms providing specific information as and when required by the consumer. Like other speakers they found the nature of online reading to be quite different from reading in print, describing it as ‘snacking’: reading less frequently and for half the time. However, online readers are more likely to take action as result of the ease of doing so.
Wayne Eadie focused on the synergies provided by advertising in both magazines and websites. Ellen Romer and Karen Ring also examined these synergies, by means of their Multi-Media Engagement survey. Overall, they found print magazines inspired greater reader engagement than magazine websites, while noting differences by genre. Like Wayne Eadie, they showed findings to support the value of cross-platform advertising.
Scott McDonald and Rebecca McPheters took a different approach by making ‘Time based comparisons of Media Effectiveness’. Their experiment was designed to measure the amount ad exposure delivered by half an hour of media consumption, comparing time spent with magazines, Internet and TV. They found the three media to differ widely in the rate and effectiveness of ad delivery. Scott and Rebecca used eye-tracking software to analyse the Internet sessions and, in particular, the response to banner advertising, finding that: “…37% of the ads served were seen either actively or passively by respondents- while their eyes stopped on 32% of the ads” Furthermore, only 11% of ads were recalled.
The final word went to Leendert van Meerem who also described an experiment using eye-tracking to measure the value of advertising in print against online. The experiment found that ads on websites have longer lasting eye fixations, which Leedert attributed to the context. However, he highlighted the tendency of more experienced online surfers to avoid ads altogether.
Developments in multi-platform delivery have refocused attention on understanding and valuing publisher brands. This theme ran throughout the conference, and was concluded in the last session with a paper from Tabu Eboreime and Marjo Martinez: ‘From number of print readers to total brand measures’. This paper looked at how a magazine’s brand image differed according to channel. For many attributes they found little difference, but noted higher scores for ‘trust’ from those reading the magazine in print.
The theme of trust was also picked up by Jennie Beck’s paper looking at the future of newspapers in a digital age. While noting the on-going relevance of the relationship of trust delivered by familiar news brands, she also highlighted how online consumption leads to “differences in perceptions of brand equity”, largely as a result of snacking on the brand rather than consuming it as a “tangible and comprehensive package”. Jennie’s paper also outlined the ways in which reading news in print is changing, with some suggestions as to how to adapt.
MRI shared an overview of the key findings for print advertising from their Starch Ad Measure studies, based on a combined sample of over 30,000 Internet access panel interviews. Marty Frankel told the conference that 20% of the variation in the propensity to note ads is explained ‘non-creative’ factors. Mickey Galin got into the ‘nitty gritty’, in her own words, with a practical look at the implications of ad position and size of ad, and the interplay between the two. She was also able to highlight some of the characteristics of high-performing ads, such as the use of recipes in food ads.
From average issue readership to advertising exposure
Several speakers emphasized the never more pressing need to bring print measurement into line with the measurement of other media. There was a call to arms from the ‘two Peters’ (Callius and Masson) in their paper ‘The War of the Media Weights’. They highlighted the implications of “major differences in what constitutes advertising reach…resulting from the way each medium is measured” in a multimedia planning environment.
With this in mind, further progress has been made in moving print along the ARF’s scale of media evaluation metrics. Several surveys have developed measures of advertising exposure, and are using them to overlay existing measures of vehicle exposure.
One way of moving forward was described early in the conference with a paper from Norway about the introduction of an ‘extended currency’. Their new survey provides an estimate of likelihood of seeing an advertisement in a given newspaper alongside the traditional measure of Average Issue Readership for the newspaper as a whole. This involves modelling Content Exposure Probabilities for readers, informed by a separate page traffic study. The paper went on to win the Best Paper award for Knut-Arne Futsaeter, Ingvar Sandvik, and Tore Ostnes.
In Pakistan an ambitious new single source survey described by Peter Masson has incorporated a measure of the probability of passing an average page for newspapers, though not as yet magazines.
Taking a different approach, Michael Hallemann and Gabriele Ritter described the development of MediaScan by ag.ma, the German readership survey. Respondents are given a hand-held scanner to register all of their reading processes by scanning in printed barcodes over a two week period. This yields data on the speed of audience accumulation and on the rate of page exposure per magazine, including the number of pages opened up several times. If the tests are successful these data will be used to overlay the standard recency data collected by ag.ma. In this way, ag.ma hopes to account better for the performance of print and improve the level of comparability with other media.
Making the case for print in media mix models
Similarly fired by the need for print to hold its own in a multi-media planning environment, two papers highlighted the importance of improving the accuracy of media mix models – one from Mark Reggimenti and Judy Vogel and one from Britta Ware and David Shiffman.
Using the Internet as a data collection tool
The ‘Getting it Right’ session started with complementary papers from Anne Crassweller and Julian Baim comparing the samples provided by different Internet access panels, and identifying the weaknesses and inconsistencies. However, as Julian pointed out, panels offer considerable opportunity if they are used to collect additional ‘relative’ data, rather than an ‘absolute’ measure of readership.
In the Netherlands part of the sample for the main readership survey does come from an access panel, though Irina Petric and Marion Appel stressed the importance of maintaining a partial random sample as the “foundation” of the survey. Their experience of using an access panel was relatively positive – something they attributed to an insistence on high quality recruitment and control of the panel.
A number of papers from Ipsos MediaCT and Kantar Media got into the detail of how best to design online surveys, looking at how to keep respondents engaged and the implications of questionnaire design on the quality and shape of the data obtained. Top tips from Liz McMahon and Rebecca Stamp included taking particular care with the first five minutes of the survey, when most drop-outs occur, and ensuring that questions are no more than ten words long.