The historic numbers behind the NYT’s October audience increase

Media | Fundraising

Thanks to aggressive pitching, my TimesSelect series got out– mostly from people I personally know. UMass-Amherst's Bill Densmore spoke to me about it for an upcoming podcast;  from Dan Gillmor: "prodigious work"; from Jonathan Dube: "very detailed analysis"; BusinessWeek Blogspotting's Stephen Baker passed it along as well.

But, as always, pitching thick research is an uphill battle. I exchanged emails with a few high-level people at the Times, but it only resulted in resolving a misstatement that Vivian Schiller had made. None of the Times bloggers I emailed (Opinionator, Freakonomics, Open Source) showed the slightest interest. Slate took a pass as well. Jim Romanesko the thought that the data was "old" and wouldn't run it (since the Buzz graphs were datelined 10/12/07). As usual, I didn't hear back from the Columbia Journalism Review. But I see that CJR is still following the story somewhat. Last Friday, Paul McLeary wrote:

Boy, the New York Times’ aborted TimesSelect experiment just keeps looking dumber and dumber. According to a post at, the online version of the Times just keeps growing. In October, attracted over seventeen million unique visitors, up from more than fourteen million in September, making it the paper’s best month ever. It’s likely no coincidence that the record fell just after the paper scuttled its pay wall in mid-September, opening the entire daily package to readers online.

Without the grace of numbers, CJR can only offer the grace of language ("dumber and dumber"). I have the numbers. I took Jon Dube's monthly NetRatings rankings over the last five years and dropped them into this chart.

Here's what we learn: saw a 19% increase over the last month. The average growth for the Times has been 1% a month over the last few years, though the month-to-month swings wildly, with a 6% average, and a 4% standard deviation. The NYT last had a 19% monthly growth in August 2005, for which I have no ready explanation. It also dropped 15% in September 2005, not returning to 13m montly readers until December 2006.  (But, truth be told, August and September of 2005 were historically high.) So, yes, is it fair to jump to the conclusion that this was a result of the end of TimesSelect readers. (On, Vivian Schiller also cited the slideshow increases — now 7% of the site.)

From the Compete data, it looks like the restored Op-Ed readers drove 80% of the increased numbers. Compete could also tell us (if we asked nicely) whether the percentage of search requests from Google grew from the 18% it was on September. By the end of the month, I noticed, Google web search started showing the 1.9 million articles from 1987-2007. (Few people are aware that the articles had been in the Google News archive for two years, but it was Google's policy to not include them in the web search purely because of the paywall– quite different from how it treats scientific journal articles.)

As for "dumber and dumber," well, that's a conclusion one can draw without access to the numbers. The advertising dollars simply weren't there at the time. And clearly the two year break did not permanently scare away the loyal-yet-not-committed readers. As for the "loss of influence" argument that some have flaunted, it's still difficult to prove. I found a January article from Radar Online be Jebediah Reed which concluded that, when it came to the pundits and Iraq, "the losing bet takes the pot." Hawks like Thomas Friedman and David Brooks are doing better than ever. The pundit buzz graph shows that as well.