Television Rip-off Points – What’s the remedy?

Guest Column: Debraj Tripathy, Marketing Communication & Advertising Consultant, and former MD of MediaCom, South Asia, enumerates four actions that he says may solve the issue in the long run

e4m by Debraj Tripathy
Updated: Oct 14, 2020 9:17 AM
Debraj Tripathy

The current media brouhaha over alleged TV rating manipulation and the variety of opinions on solutions has left many advertisers, and the general public confused. Given the volume of discussion already published/ aired across media, this article will skip the modus operandi and address the question of remedies… What can be done to prevent, or more practically, reduce the scale and frequency?

Currently, whether the crime can be confirmed or not, the issue of panel tampering is very real. Maintaining the confidentiality of the identities of the panel households is no easy task. There is enough incentive for a few unscrupulous broadcasters, past, and current employees of research agencies, to breach confidentiality protocols and manipulate viewing patterns of panel households. Only 44,000 people meters, spread geographically across the country, are used to arrive at viewership estimates of close to 850 million individuals in 200 million TV households. So, manipulating just a few households, say, less than 8 to 10 households in any market, enables one to swing TV ratings any way one wants.

Many solutions have been proffered… from using mobile phones to track viewership (maybe doable, but not easy), to increasing the panel size by installing more people meters in more households (doable, but expensive), to banning TV ratings and buying TV inventory basis subjective/ soft metrics. While I don’t disagree vehemently with many of these, some are impractical. Here are four actions that may solve this in the long run, not completely ever, since crooks will always find a way to beat the system, but to a significant extent.

Action 1 (Immediate): Business pressure from advertisers and their media agencies

Advertisers and broadcasters who play by the rules suffer huge losses. Advertisers lose scarce advertising resources, chasing imaginary consumers. Honest broadcasters lose revenue that should rightfully be theirs. For every instance that is detected, many may be going undetected, which makes it difficult to estimate the real degree of loss.

Key advertisers, either singly or as a body, will need to take a strong stand on how to deal with, maybe blacklist, erring broadcasters for a period of time. There are examples where this has worked – P&G’s strong stand against opaque and convoluted digital media supply chains resulted in subsequent action by other large advertisers and had an impact on the level of transparency in the digital media domain.

Action 2 (Short to medium term): Employ Machine Learning algorithms to enhance the surveillance capability

It is humanely impossible to scan the voluminous data that BARC collects to detect fraudulent practices. So aggregated data is examined to see if there are discrepancies. This results in loss of granularity, and hence detecting anomalous changes at household/ individual/ time band levels becomes difficult, more so, when a smart crook employs, what I term, creeping, and distributed manipulation. In creeping, the manipulation is done gradually in bits, over an extended period, say 6 to 8 months. It starts with say, manipulating 5 minutes in a day in week 1 to gradually building up to 3 hours in week 32. Distributed manipulation is when one infiltrates 10 panel households when only 3 are required and then actively manipulates 3 at random. When both are done in tandem, the slow growth in viewership looks like a gradual natural change. Such manipulation, even once detected, makes it difficult to pinpoint the source of the anomaly.

Given the volume of data that BARC has AI algorithms can reliably correlate changes in viewership/ ratings with features like changes in content, events, time of the year, changes in seasons, changes in internet search patterns, changes in social media sentiments, etc. Clustering algorithms can detect market-level changes in panel household clusters. These will help in robust surveillance and in detecting sudden, creeping, and distributed manipulation of viewership.

Action 3 (Long term): Enhance the viewership data tracking process through technology

This is the difficult bit with probably no immediate solution.

As suggested by many, the most obvious way to stop manipulation is by increasing the number of people meters (panel households). Large numbers are more complicated and costlier to manipulate. The incentive thus goes down, though it might still be attractive to a few. But, who’s to bear the cost? Deploying and maintaining people meters is not cheap. Additionally, increasing them does not necessarily result in more significantly accurate data. Given the strains on the cost issue that the industry went through when BARC was launched, we may not be able to bear the huge cost of additional people meters.

BARC has an elegant solution to this already (though it will take time to see the light of day) – tracking viewership habits through set-top-boxes (STBs) of various cable and DTH players using return path data. But, a majority of the STBs in India do not have built-in two-way communication capabilities. So, finding a representative sample will be slow and involved. Despite this, the idea of tracking viewership through STBs is probably the best solution we have at the moment. It is inexpensive (no cost of people meters) and is attached to a TV set, so is as reliable as it can get. STBs can then be used to increase the number of panel households, manifold.

Other solutions have been suggested like tracking viewership through mobile phones. Technologically this is possible, and companies are doing this on the sly, at the moment. They listen to whatever is happening around one, through the phone, and track it. However, even if we were to solve the massive problem of getting authenticated profile data (which advertisers use to target consumers) for a large mobile phone panel, there is virtually no guarantee that one will carry one’s phone around while watching TV.

However, the aim is not just to increase the number of panel households to be tracked. The aim is to have enough redundancy in the panel so that one may choose random households (of course keeping population representation in mind) to include in the day/ week’s viewership measurement. The STB approach or a practical mobile phone approach will enable this to happen at relatively low costs. To clarify, the suggestion is that we track 1 million panel households but, pick only 50,000 of these households in a day/ week (assuming 50,000 is enough to give us statistically significant results) at random to determine viewership.

Action 4 (Immediate): Open up viewership data to outside experts for examination/ scrutiny

Making anonymized raw data available to outside experts for scrutiny and examination will enable deeper and distributed examination/ scrutiny of the patterns of viewership. A reward programme, like the ones that Google or Facebook runs to find bugs in their code, would help in getting a large number of outside experts to examine the data and find anomalies. In cases like this, a thousand heads are always better than a few wise men trawling through tera-bytes of data.

Malpractice is not new where huge sums of money are involved. For instance, agencies, platforms, and advertisers continue to grapple with manipulation in the digital space, which could be to the tune of 45% - 50% (more about this in a later piece). In India, fraud in TV viewership just gets more eyeballs because people understand TV advertising better than digital advertising.

It might not be possible to eliminate illegal practices completely but reducing it in scale and frequency is an objective worth pursuing.

Disclaimer: The views expressed here are solely those of the author and do not in any way represent the views of

For more updates, be socially connected with us on
WhatsApp, Instagram, LinkedIn, Twitter, Facebook & Youtube