• by Joe Mandese  @mp_joemandese, November 21, 2012

Nielsen disclosed more about the magnitude of Superstorm Sandy on national and local TV ratings during a “recovery plan” briefing with clients Tuesday morning, confirming that at least eight metered markets representing nearly 17% of U.S. TV households were impacted to some degree, and that the method it used to fill in the blanks for national audience estimates -- mathematical weighting -- “overstated” some national TV ratings during the period.

“I’ve been looking at disasters, if you will, in terms of their impact on television, as long as anyone has, and in terms of recovery, I think this honestly is the best we’ve ever had,” long-time Nielsen executive Pat McDonough told MediaDailyNews following the recovery webinar with clients, which she presided over.

McDonough confirmed that Sandy most likely was the biggest natural disaster in terms of overall impact on Nielsen’s local and national systems, although she said the long-term effects were not nearly as devastating as the effects of Hurricane Katrina on New Orleans, whose population fell so much that it dropped Nielsen’s market ranking for the city by 10 positions in its aftermath.

By contrast, McDonough said the long-term effects of Sandy will be relatively small. In the New York market, the largest and most severely impacted of all of Nielsen’s metered markets, she said only 21 of the 1,000 households in its sample were permanently destroyed by the storm.

“It’s about a little over 2%,” she noted, adding that Nielsen is working quickly to replace those households with alternate homes to bring its New York sample back to 1,000.

While New York, which represents 6.468% of the nation’s viewing, was most affected -- so much so that Nielsen has actually withheld its official ratings estimates for the entire month due to incomplete data reported for three of the four weeks of the November survey period -- she confirmed that eight metered markets ranking from Philadelphia to Providence did not report data for at least one day during the storm. Philadelphia, the next-biggest market affected after New York, lost five days of data, and the other markets lost one or two days.

McDonough said some clients would continue to receive some daily data in the impacted markets via various electronic data delivery systems or third-party processors, and it was up to TV buyers and sellers to decide what to do with the data.

“All of that data is available for commerce. It’s up to the market to decide what to do with it,” she said, adding that Nielsen will be footnoting all of its official reports and data tapes with a “breakout indicator” noting that the numbers may have been affected by “atypical TV viewing patterns” during the periods. She said it was only the second time Nielsen ever did that. The first time was following the Northeast blackout of 2004.

McDonough said Nielsen executives were pleased to see that their systems and metering equipment in sample households worked properly once power was restored to those homes, but said it still isn’t clear exactly how many of them might have been reporting data due to back-up household generators but were otherwise unable to watch television.

On the flip side, she said some TV viewing behaviors might have been increased in other situations, noting that she had power and TV reception in her home and played host to visiting relatives and friends who increased viewing in her home during the storm because of that.

She also noted that households relying on either over-the-air broadcast signals or satellite TV subscription services were faster to come back on Nielsen’s grid than cable and telco service providers, which also had to deal with downed lines and other hardware issues.

She said none of the markets were completely knocked out at any given point in time, and that even in New York, the number of households reporting data never dipped below 450 homes out of its 1,000 household sample.

That said, she did acknowledge that the overall effect, including an undisclosed number of national sample households outside the eight metered markets affected, did impact national TV ratings estimates during the storm.

However, she said the “in-tabs,” a Nielsen term for the percentage of households reporting useable data on a daily basis, never fell below its minimum guidelines for reporting national ratings estimates.

“It dropped to 89% during the very worst of it,” she said, noting that this was about six percentage points lower than Nielsen’s typical daily in-tab of about 95% of national sample households reporting data. Nielsen’s threshold for withholding national TV ratings estimates is an in-tab of 80% or lower.

“Overall, we took about a 6% decrease in terms of the total country,” she said.

McDonough said Nielsen’s method for estimating national ratings during that period – mathematically “weighting” homes reporting data in the sample to represent the homes that were not able to report – did “overstate” national TV estimates by some unknown order of magnitude, but she said that method is Nielsen’s normal procedure.

Ultimately, she said it’s up to the marketplace to decide how to treat those ratings in terms of advertising deals and ratings guarantees.

The days affected by the storm included some of the most expensive TV advertising buys of the year in the national and local TV advertising marketplaces.