Rentrak 'Big Data' Tracks Pre-Hurricane Sandy Stats

Hurricane-Sandy-AA2Rentrak, a growing TV measurement service, says viewing just before Hurricane Sandy sparked stronger-than-usual daytime viewing in the New York market -- only to drop well below averages during prime time when the storm hit.

Surveying its some 90,000 homes in its sample, Rentrak says the homes using television (HUTs) on the day of the storm -- Oct. 29 -- were at a 50% level at 9 p.m., rising to 55% at 5:15 p.m. on Oct. 29.

This compares with the week before on October 22, with daytime homes-using-television viewing of 39% at 9 p.m.  dropping to around 36% for most of the day -- then rising to around 55% at 5:15 p.m., the beginning time of the usual steadily growing higher TV usage.

But on the day of the storm, after around 5:15 p.m. -- when the storm would have a major effect and knock out electricity to some 8 million homes in the region -- TV usage remained around the 50% mark until 9 p.m., far lower than under normal conditions.

By way of comparison, the week before -- a normal viewing day -- witnessed sharp, rising HUTs growing to 97% by 9 p.m.

Bruce Goerlich, chief research officer of Rentrak, stated: "Rentrak's 'big data' approach means that we can continue to report and avoid going black, even in areas hit by extreme weather. Our dependability is a huge differentiator for our clients."

For its part, Nielsen -- which local and national advertisers depend on for current ratings data for their media buys -- said the storm so severely disrupted its viewing collection operations that it would not be releasing complete TV data for November for the New York market. A number of major East Coast markets could also be affected.

HH-Using-tv-B

Recommend (7) Print RSS
6 comments about "Rentrak 'Big Data' Tracks Pre-Hurricane Sandy Stats".
  1. Darrin Stephens from McMann & Tate , November 19, 2012 at 11:39 a.m.
    Of course, Rentrak has no way of knowing what percentage of homes stopped watching tv because of a power failure or just turned the set off. Actually, Rentrak really doesn't even know when the set's off because they only measure cable boxes, not TVs. They have to guess.
  2. Ernest Smyth from None , November 19, 2012 at 7:47 p.m.
    Rentrak doesn't guess. It takes the information it does have, and models for what it doesn't. Very similar to what Obama's campaign did in its use of big data.
  3. John Grono from GAP Research , November 19, 2012 at 8:10 p.m.
    Ernest, I think Darrin's point is not modelling data that you don't have, but validating data that you do have. For example, I leave my STB on 24/7 tuned to whatever the last channel I watched was. Unless that is rigorously edited out then the data is over-stated. Also, how are people per set and demographic profiles generated. Darrin, it is pretty common for the on/off status to be monitored on the STB. If you get either an On or Off you know it has power - and you know the status of the box.
  4. Nicholas Schiavone from Nicholas P. Schiavone, LLC , November 19, 2012 at 11:57 p.m.
    What the hell are Wayne Friedman and Bruce Goerlich talking about? Is Wayne Friedman actually a Time Lord? Is Bruce Goerlich now a PR Agent? Mr. Friedman's sense of chronology is mind-numbing to this reader. Who talks about audiences "rising" -- as they move from 9 PM to 5:15 PM? Who talks about the "daytime" HUT levels -- at 9 PM? And Mr. Goerlich, who for years could not get close enough to Nielsen Management to obtain yet another proprietary new-business research weapon for his agency, now cites Rentrak's "big data" approach as superior to Nielsen's MRC-accredited statistical sampling approach because 'big data' systems can "avoid going black." To me, it sounds like the lights are still out at Rentrak, if Mr. Goerlich can imply that the nonrepresentative Rentrak "sample" is seemingly impervious to power outages, which gives it a definitive advantage over Nielsen's first-rate area probability samples that warrant no mysterious modeling. At this point, it would seem prudent for Rentrak to consider getting beyond the TV ratings business and the electrical grid and stick with the "natural gas" that fuels Mr. Goerlich's hyperbole. Last week, I recommended -- in response to a Joe Mandese article containing industry criticism of Nielsen's storm handling -- that instead of complaining, the industry would be well-served by inviting an MRC-led review of the major audience research companies' approaches to emergency management. While Nielsen's silence is preferable to Rentrak's opportunism, both vendors and their customers need to understand what happened to their systems during and after those unprecendented natural disasters, in order to determine how operations can be improved. PS Darrin's comments are spot-on in his critical assessment of the situation. Unfortunately, Ernest appears to have been temporarily overcome by the "natural gas" associated with Rentrak's press release. Cheers! (http://www.prnewswire.com/news-releases/rentrak-ratings-during-superstorm-sandy-show-overall-new-york-city-tv-viewing-spike-179937361.html)
  5. Rob Frydlewicz from RAF Consulting , November 20, 2012 at 1:42 p.m.
    A HUT of 97% at 9PM is "normal"? Based on my 30 years in the research biz, typical HUTS at that peak time of viewing fall in the 65-68% range. Unless Rentrak uses a different definition of HUT.
  6. Nicholas Schiavone from Nicholas P. Schiavone, LLC , November 27, 2012 at 9:56 p.m.
    The Juily Ben comment is a travesty. MediaPost Publications ought to investigate and sanction the parties responsible for this abuse of a valued digital publisher - and its valuable readers.