Don't trust the new Indian Readership Survey (IRS), it is full of shocking anomalies."
That is the message that the Indian Newspaper Society, or INS, sent out through advertisements, press releases and articles in the papers owned by its powerful roster of members all of this week and last.
This includes The Times of India, Anandabazar Patrika and Dainik Bhaskar, among others. It is a strong message. But it has not yet rattled the Media Research Users Council, or MRUC, which owns IRS. "If you find a real error in the study, we will fix it," says MRUC Chairman Ravi Rao, who is also leader (south Asia) at media buying agency Mindshare.
The survey, which was carried out by Nielsen India on behalf of MRUC, has drawn Rao and MRUC to the centre of a feud that threatens to destabilise the order in the Rs 22,400-crore print media. IRS is one of the key metrics used to buy and sell advertising space in newspapers and magazines. Last week, MRUC released the results of a re-hauled and redone IRS. It has upsets across states, languages and cities. Plus, there are "anomalies."
Newspaper publishers first went into a sulk. Then, earlier this week, they went ballistic, demanding the withdrawal of the data within 24 hours. When MRUC said that the newspapers would have to wait till February 19 for a decision, INS advised its members to withdraw from IRS. Several have done so already. Magazine publishers are threatening to follow suit. "It is a war of nerves," says one MRUC board member.
"Until and unless it provides a solution (for the data anomalies), we cannot move forward. As of now, this survey is over," says Girish Agarwal, director, Dainik Bhaskar Group. Adds D D Purkayastha, CEO, Anandabazar Patrika, "This data is completely haywire and highly inconsistent."
For many, the survey results are hard to fathom. "In Kerala doctors don't read Mathrubhumi and Malayala Manorama, illiterate people read the newspaper," says Jayant Mathew, director, Malayala Manorama, rattling off the aberrations in his home state. Others point to The Hindu Business Line, a Chennai-based newspaper that has more readers in Manipur than in Chennai!
The state-wise upsets are "interesting." Rajasthan Patrika is ahead of Dainik Bhaskar in the latter's home state Madhya Pradesh; and in Uttar Pradesh, a Dainik Jagran stronghold, Hindustan has made huge gains.
"Many of these papers have lost readership on their home ground," says another MRUC board member. "I don't think the decrease in numbers matters as much as the errors," argues Mathew. "Why would a publication which has gained because of the survey condemn it," asks Agarwal. Divya Bhaskar, the group's Gujarati paper, has become the number one in Gujarat. "In circulation, in urban Gujarat, we are ahead of competition; however, in rural Gujarat, the competition has an edge," says Agarwal.
What really is the issue?
One is cherry-picking of data. While at the national level, the sample is a large one at 235,000, it does not translate into enough representation from each city or each language or each type of reader. So looking for say doctors in Kerala may throw up funny results because there may not have been enough doctors in the sample to adequately represent their population.
The IRS terms of usage states that "users must not analyse or report numbers (especially average issue readership) on an unweighted respondent count of less than 30." It is when users do not follow this rule and drill into smaller, statistically unstable clusters that anomalies crop up. "This way I can find a hundred anomalies in the 2012 data too," says one MRUC member.
If the national sample was say 500,000 or 700,000 then digging deeper into clusters would have been possible, says one statistician.
Two, "from day one we have been saying don't compare, but INS didn't want to listen," says Rao. The advisory issued by MRUC states that this survey cannot be compared to previous ones because everything in this IRS has changed - from the way the interviews are done to how data is collected. Usually, the population for every round of IRS is estimated on the last available census.
Till 2012, it was was based on estimates derived from the 2001 census. Now that the actual numbers of Census 2011 have come in, many things have changed - city boundaries have been redefined and population has dropped in several metros. The biggest impact this has had is on the estimate of the total number of print media readers - they have fallen from 353 million in 2012 to 281 million in 2013.
In the late 1990s, "when the ratings system changed from diary to people meter, there was a 20 per cent drop in several genres. And we used to say don't compare diary to people meter; similarly, it makes no sense to compare this data to the previous one," says one member.
Maybe MRUC has a point. But some of the numbers are strange enough to have alerted the team before releasing the data. For example, how did all newspapers in Andhra Pradesh lose 30 to 60 per cent of their readership? A new methodology was bound to throw up upsets. Could MRUC have worked at making the transition from a system the industry had been using for 17 years to a new high-tech one smoother?
This could have been done through workshops where users could explore the data and also be warned of the inconsistencies possible if they did not follow the terms of usage. Rao says, "We made a full-fledged presentation to all the agencies and users, and we did roadshows in Mumbai, Delhi and Bangalore."
Similarly, why were publishers not better aware of the changes that were taking place? The current board of governors of MRUC includes senior managers from five major print firms. The technical committee which worked on the new IRS has 20 members, six of them from print firms (including one from this paper). While some publishers agree that they should have been better clued in, another one says, "Whenever we had queries, they threw jargon at us."
Clearly, one big problem is the communication between MRUC and one of its biggest user members - the print media.
The way out
"Once this survey is out, it is not just opinion, it is gospel (for media buyers and advertisers)," says Agarwal. He is right. Any fall or rise will be used as a tool to negotiate ad rates. And to avoid that INS had asked MRUC to look into the flaws. However, with both parties sticking to their guns, what will happen is anyone's guess.
If MRUC holds back the data, runs through two to three rounds and irons out flaws if any, it could help. It also gives users time to get used to the new system. MRUC keeps saying this data has to be looked at over a period of time. So maybe a few more rounds will help users appreciate the strengths of the new methodology.
If publishers keep pulling out of IRS, it causes several peculiar problems. The brands that have withdrawn, say a Dainik Bhaskar or Anandabazar Patrika, have also directed MRUC that it cannot include or use their mastheads while doing the survey. With several of them gone, will the survey have the same validity?
And what will the publishers do? "We don't want to be in a position where there is no currency. We don't want to go the (TV) ratings way," says Mathew. There is some talk of having their own metric. "The structure has to be changed. We need to be in better control (of the metric)," says Purkayastha.
That is a tricky one. "Globally, the best known audience measurement is from joint industry bodies," says Paritosh Joshi, chairman of the IRS technical committee. For instance, the Broadcasters Audience Research Board, an industry body, does TV ratings in the UK. Why would INS, whose members make up for roughly 40 per cent of MRUC, break away from it and invest in a new metric that most advertisers may not take seriously?
Any break therefore would harm both IRS and the publishers. MRUC had asked INS to wait till February 19 when the Readership Studies Council of India, the body that now governs IRS, meets (see box). Hopefully, by then, tempers will cool down somewhat and everybody will be in the mood to talk.