Page Menu Options

FIPP: What’s new in readership research worldwide?


What have been the principal developments during the last two years among national readership surveys around the world? Guy Consterdine, FIPP’s insight consultant, has been examining a newly-released review of 103 national readership surveys in 78 countries, and 58 specialist audience surveys (for instance among business people).

The author of the review is Katherine Page, media research consultant, who was commissioned by the Print & Digital Research Forum (PDRF).

Integrating print and digital audience estimates

Katherine Page found that the principal developments during the last two years have been a move towards integrating print and digital audience estimates into a single database, and trying to keep pace with the increasing number of platforms and devices used to access publisher content. The aim has been to provide publishers with a measurement of total brand reach across all platforms, and the ability to carry out detailed analyses of advertising campaigns, through reach and frequency planning software.

Most of the 161 surveys do now include some measure of digital reading alongside and in combination with print readership estimates. That said, there is a great variety of ways in which this is achieved, some more advanced than others.

The largest group of surveys is those which rely on respondents’ recall of their visits to publishers’ websites. This can be a useful starting point for having some information put into the market on digital reading of publisher brands, but as Page commented “it is difficult for participants to remember (or even know) the full scope of their digital activity, particularly occasional or one-off visits”.

A second group of surveys, 21 in all, are using sophisticated data integration techniques to combine the print readership study with external data on website audiences measured automatically by passive methods. Thus the external data does not rely on human memory but records actual behaviour. Common sources for the external website data are comScore and Nielsen, but there are others too such as TGI Clickstream – for example the China National Resident Survey (CNRS) integrates TGI Clickstream digital data into the print survey. In Australia the EMMA survey uses fusion methods to incorporate Nielsen website audience information into the published resource. Among other surveys which use various integration methods are the Estudio Nacional de Lectoria in Ecuador, Consumer Media View in Thailand, SNA Focus in Romania, Media Atlas China in China, and Peru’s Estudio Nacional de Lectoria.

These fusion or calibration methods are a big advance. However they are not without limitations. As Page writes, “a particular concern is the ability to reflect adequately the duplication of readership between print and digital platforms”. If the print and digital duplications which emerge from the fusions are not realistic, the estimates of total brand footprint will not be realistic. I have been conscious of this myself in my work in Britain with the National Readership Survey’s PADD (Print And Digital Database). For magazines, especially those with niche audiences, producing credible duplication patterns has been a real concern.

This leads on to the third, newest and smallest group of surveys: those designed specifically to facilitate integration of print and online reading. These have a single-source component in which some respondents have their internet surfing measured passively (instead of recall) as well as answering questions on their print reading; this yields very high-quality duplication data between the platforms.

The few examples include the MACH Basic study in Switzerland and The One survey in France. What I am most familiar with is the forthcoming AMP (Audience Measurement for Publishers) survey in Britain, with which I am working. To be launched in 2016, there will be three interlocking principal elements:

  • 35,000 face to face interviews collecting print readership and other data;
  • a 5,000 respondent single-source passive panel, recruited from among the 35,000, whose visits to publisher websites will be recorded through software on all their devices; this will yield excellent duplication data between the print audience and digital audience for each brand (subject to sample sizes);
  • comScore website audience data which will be integrated into the system, utilising the duplication data from AMP’s own passive panel.

I anticipate that more national audience surveys will evolve into this third group in the coming years, the main constraint being the cost. It must be seen as a vital investment in the future of content publishing.

Digital editions

An increasing number of surveys are now reporting audiences to digital editions of magazines and newspapers, either on their own or built into the estimates for the print edition. It has to be said that more newspaper than magazine digital editions are covered because of the generally larger audiences of the newspaper products, making them more practical to measure on grounds of having sufficient sample size.

For example, in Spain the EGM survey (Estudio General de Medios) is now measuring digital editions for all print brands. Among the surveys reporting only on newspaper digital editions are Norway’s Forbruker & Media, Ecuador’s Estudio de Lectoria, Italy’s Audipress, and France’s The One survey.

Increase in online interviewing

Another trend identified in the Katherine Page review was an increase in online interviewing during the last two years. A number of national readership surveys have been relaunching or undergoing major change in which there is a switch from personal interviewing to online interviews – either totally, or in part. Examples include the MA Media Analyse in Austria, the Print & Digital Audience Measurement in Canada, the Roy Morgan Single Source in New Zealand, and two surveys in Slovenia – the NRB and the Target Group Index.

Nevertheless it is still a minority of national readership surveys which use significant proportions of online interviewing (about a fifth of surveys), compared with personal interviews accounting for about 60%. Online interviewing is, in general, less expensive than personal face to face interviews, but the big problem which online methods must overcome is to recruit a truly representative sample.

Measuring specific issue readership

The great majority of national readership surveys measure reading by means of the Recent Reading method – that is, questions of the general form “When did you last read a copy of…?”

However, there are now four surveys which currently use or are about to use the Specific Issue Readership (SIR) method for magazines. With this, respondents are shown the front covers of the most recent issues of the magazine (for example, the last six issues), and are asked which if any of those issues they have read. There are several arguments in favour of this method, and several against, but one of the most significant advantages in my view is that SIR automatically provides a realistic measure of the rate at which readership of an issue accumulates through time. In most readership surveys this is a missing element, but readership accumulation should be a vital part of planning advertising campaigns and demonstrating the effectiveness of magazine advertising.

The four studies using SIR are Magascene in Japan, EMMA in Australia, NOM in The Netherlands, and the forthcoming AMP in Britain.

Some country by country developments

Few of the 161 surveys in Katherine Page’s review are standing still. The majority are introducing new developments of some kind, as publishers respond to several pressures: from advertisers to prove the size and nature of print audiences, and the value of print advertising; demands for higher technical standards in audience measurement; or in response to the advance of digital reading.

Here is a taste of the variety of developments in the last two years:

  • Afghanistan: Ipsos are launching a readership study which also includes questions on products and services.
  • China: the China Marketing & Media Study (CMMS) has extended the number of cities covered from 46 to 62, including some rural cities. It has also begun integrating TV ratings data into the CMMS database, and plans to integrate digital audience data in the near future.
  • Costa Rica: the Estudio Multimedios sample is now controlled by quotas, and the questionnaire has been harmonised with the Multimedios survey in Panama.
  • Czech Republic: the Media Projekt survey has introduced a quota sample, and increased the sample size from 2,800 to 3,500. The TGI Market & Media Lifestyle study has increased the proportion of interviews conducted by online interviewing, from 17 per cent to 25 per cent.
  • Finland: the KMT (Kansallinen Media Tutkimus) now measures magazine readership by online interviewing, while newspaper readership is measured by telephone interviewing (as magazines used to be). Magazine publishers preferred online interviewing because they judged that the ability to show logos and front covers would improve the accuracy of the results.
  • Germany: the Mediascan study has been launched alongside the main MA (Media Analyse) currency, to provide measures of ad page exposure, number of pick-ups, and time-based performance metrics. To collect the Mediascan data, a panel of 2,000 respondents are given devices to scan their publications every time they read.
  • India: the Target Group Index sample size has been increased to 40,000.
  • Luxembourg: the Etude Plurimedia has introduced a mixed methodology by adding 1,000 online access panel interviews to the existing 3,000 telephone interviews.
  • Mexico: The Estudio General de Medios (EGM) has introduced a 24-hour past-day electronic diary. In addition some interviews are now conducted online.
  • New Zealand: Nielsen’s Consumer & Media Insights (CMI) study is testing potential online completion of the survey, with results pending.
  • Peru: the Estudio Nacional de Lectoria is now administered by tablet. Previously a mobile phone was used.
  • Romania: the SNA Focus survey has begun publishing ‘total brand’ audience estimates, combining print, the website, the Facebook page, and any other ‘brand extension’. Moreover the universe has been expanded to the whole population aged 14-17, from being just urban areas. To achieve this, the sample size has been increased from 10,000 to 16,000.
  • South Africa: the four largest publishers have formed a joint industry body, the Publisher Research Council, to plan, test and commission major new developments, including multi-platform engagement.
  • Vietnam: the Media Habits Survey (MHS), which covers the four major cities, has doubled its sample size, partly in order to publish results more promptly. Respondents are also re-contacted for an interview about their product consumption.

Readership research continues to adapt and evolve in response to the changing world of publishing. It is notable that these developments are speeded up by the exchange of information between practitioners and between countries – an exchange facilitated by organisations such as the Print & Digital Research Forum, FIPP, and magazine national associations (among others). The Katherine Page review every two years is a significant element in this productive process.