I don't know how frequently I've been wrong as a journalist, but I'm certain it's more than I'd like. It's not because I don't have the skills or work hard to avoid them, but because what I do isn't always an exact science and depends on the quality of the information I have when I write and publish a story. But I'm human, and so are most of my sources, and sometimes we make mistakes. When we do, a journalist has an obligation to set the record straight and publish a correction.
I now know that a story I wrote last week about a Nielsen ratings glitch, also contained a big glitch. It incorrectly attributed the potential impact of a glitch in a Nielsen's system for processing audience frequency estimates to the system Nielsen uses to generate its C3 ratings, which are the currency of the national TV advertising marketplace, and the basis of most advertisers' audience guarantees. I tried to make that error clear in a story we published on Friday based on a report we received from Deutsche Bank, in response to Nielsen's shares taking a nosedive after our first story - and some related news made by Mark Cuban - came out.
I don't know how frequently Nielsen makes mistakes, but I'm also quite certain it's more often than some of the company's stakeholders would like, because I usually only learn about it when one of them lets me know.
Last Wednesday, one of them passed along a copy of a notice Nielsen sent to clients disclosing that seven months of data produced by its NPower system had inflated audience frequency estimates. That source was concerned that the inaccurate data could have important implications, including an effect on the TV advertising marketplace.
The next source I contacted was my press contact at Nielsen, who said no one would be able to speak to me in the 13-hours before I would publish a story, and no Nielsen executive was available to speak to me until Friday afternoon.
So I and some of my colleagues at MediaDailyNews spoke to as many Nielsen clients as we could reach, some of whom said they believed the glitch had the potential to impact network audience guarantees and makegoods. That's what we published.
In response to that story, I learned from other sources, including Deutsche Bank analyst Matt Chesler, that that was not the case. So we published the second story on Friday, based on that information and linked it to the original story.
It frequently surprises me how upset people sometimes get when we make mistakes. We work hard to avoid them, and we work equally hard to correct them when we make them. So if it's still unclear to anyone who might have read any of our coverage about Nielsen inflating seven months of audience frequency estimates in its NPower system, the fact is that data has no direct impact on the ratings networks use to guarantee audience estimates to advertisers.
What remains unclear is what the impact of Nielsen's faulty audience frequency data actually is, what caused it to happen, why it wasn't detected sooner, and why, if as Nielsen is telling me now, hardly anyone uses it, the company still continues to produce it.
In its original statement, Nielsen said it is still investigating the "root cause" of the glitch, which was brought to its attention in the first place by its clients, some of whom told me they now have systems that monitor audience fluctuations with "greater fidelity" than Nielsen can. Which, of course, raises yet other questions.
If anyone knows the answer to those questions, please let me know and I will make every effort to report them, as frequently as that needs to be.
In the meantime, Nielsen has published something of its own, a statement asserting that faulty audience frequency data it produced for seven months has had no effect on the TV advertising marketplace. While it is true that the frequency data is not directly connected to C3 ratings that are the basis for network advertising buys, it's unclear what role audience frequency data plays in influencing TV advertising buys. At one time, frequency was considered a cornerstone metric used by advertisers and agencies to decide where to put their advertising budgets in the first place. For a number of years, frequency has fallen out of favor as a planning metric, but is still baked into a variety of systems, including third-party media planning software providers such as IMS and Telmar.
In the online media business, frequency still is considered a key metric, especially in highly efficient real-time bidding systems, where "frequency caps," are an integral part of a buy, ensuring that advertisers do not pay for more audience exposures than they need to.
One question that has an answer is why Dan Rather's boss Mark Cuban, a savvy technology investor who owns and operates cable and satellite TV network HDnet raised his stake in Nielsen rival Rentrak last week.
"In my opinion," he explained, "it's the best source of audience information and the future of audience measurement on both a local and national basis."