NEWS13 October 2014

Nielsen admits TV ratings errors

Features News North America

US — Nielsen has admitted that it has been reporting inaccurate ratings for the US broadcast networks for the last seven months.

According to a story in the New York Times, the error benefited ABC – the network saw its ratings improve – while all other networks were affected negatively. In a press release from Nielsen regarding the incorrect attributions, the company linked the issue to a “technical error” that arose in March of this year.

Nielsen claims that the discrepancy in viewing totals was related to “difficult-to-attribute content called ‘all other tuning with code’”, which represents between 0.1% and 0.25% of the viewing totals. The company says it has deployed a software fix to address the problem and will reprocess all impacted data going back to August 18, to be reissued by the end of this month.



8 years ago

This is no computer "glitch." It is an industry fiasco. This is about a major, persistent Nielsen malfunction, symptomatic of systemic error and much worse to follow... The real issue is not TV ratings for News by talent and network, but Nielsen Audience Ratings for TV overall: program by program, newscast by newscast. The Saturday (10.11.14) NYT's headline read: "TV Ratings by Nielsen Had Errors for Months." Hence, Television Journalists and Journalists who cover Television (and Media) should be focusing on the accuracy and reliability of the entire medium's measurement. That is the epic story here for all concerned. Just observe: Beloved Nielsen SVP, Pat McDonough, who is soon to retire, has been forced to defend the TV ratings "gin mill," while inscrutable CEO, Mitch Barns, remains invisible and silent, like Kim Jong-un. What is Nielsen hiding? It's time for the FTC to confront the ratings monopoly and force the issue of professional scientific accountability. Enough is enough. For example, Nielsen is now working on to implement plans to increase the apparent "statistical reliability" of its ratings by appearing to significantly reduce the standard errors (SE) of its estimates. In fact, Nielsen will be diminishing the "statistical validity" (i.e., accuracy) of its ratings estimates. Instead of tabulating (i.e., counting) the actual viewership from NPM (Nielsen People Meter) Households, as it ought to -- and has claimed to for more than 5 decades, Nielsen will model (i.e., arithmetically guess) the sample panel's viewing data in a substantial portion of its newly "Expanded Sample Frame." And it ought to be noted that the expansion will come predominantly from households equipped solely with "archaic" set meters, not modernized People Meters (i.e., GTAM/STPM). Concerns about this plan from respected researchers and media & marketing professionals have nothing to do with a resistance to science or advanced analytical techniques applied to big data. Rather, in its efforts to be all things to all potential customers, Nielsen is destroying the core values of its company and its products. Arthur C. Nielsen admonished: "Watch every detail that affects the accuracy of your work." Alas, the institutional memory is being rooted out for the sake of popularity and profitability. No pursuit of the scientific method or measurement principles here. If ratings are to be considered the currency of the marketing, advertising and media industries, then Nielsen is called to the highest level of accountability, from which it has fallen far short. What industry executives and decision-makers seek is not miracles, but practical wisdom. This is not a new age concept, but a principle that goes back to Aristotle, who would advise all that we must not cease from searching for the right ways to do the right things. [Nicholas P. Schiavone: Former NBC Research Chief & Chairman, CONTAM, 1990 to 1999]

Like Report

8 years ago

woah Nicholas; they got it wrong, they admitted it, they are fixing it. no need to go ape on the issue!

Like Report