Update, September 20: The Washington Post tells me that it prefers to stick to Nielsen statistics in public. I’m checking to see how easily available the stats are to the world at large, and if Nielsen can share any server-based numbers from the Post. I’ll also do other follow-up. – D.R.
Should news sites hide authoritative Web traffic stats from the public, perhaps for “competitive” reasons? No, actually. In openness on the Quantcast measurement site, however, the august Washington Post lags behind the Washington Examiner and even the rickety Washington Times.
I raised the disclosure issue in an update at the bottom of my last post, headlined TBD hyperlocal news experiment in D.C. area: Still iffy, with mixed numbers from the Alexa Web measurement service. As I see it, all major news sites, even those unaffiliated with paper newspapers and magazines, should volunteer to be subject to ABC-level verification, whatever the verification organization involved—and should put the numbers on the Web for the whole world to see. Alexa-style services are poor substitutes since external measurements aren’t nearly as accurate as those from the news sites’ own servers.
If certain sites hide stats from the public? Then some advertisers might use this to justify their bias against the online news media. Detailed, completely public stats will be perceived as more bulletproof than hidden ones, even those which advertising prospects can request. This is serious business. One Philadelphia publisher was even sentenced to house arrest for five months and three years probation and fined $20,000 for falsified circulation-related information involving printed magazines. I’m not accusing anyone mentioned here of similar conduct. I just believe that detailed online traffic information should be genuinely public, especially since competitors might be able to get the information anyway via advertiser friends.
Meanwhile you can see the problem first-hand through Quantcast, which will let you hide your stats if you register or else will let you publicly share the numbers, generated via the scripts which your techies can embed on your servers. For unregistered sites, without scripts embedded, Quantcast tries to offer Alexa-type estimates.
Here’s the Quantcast lowdown for TBD, the New York Times, an NYT developers site, the Washington Post, the Washington Examiner, the Washington Times, Gawker and solomonscandals.com.
—TBD is getting “Quantified,” which means that its Web pages will carry the computer code through which Quantcast will measure traffic. Will this information be truly public? I don’t know. It won’t be, however, if TBD is as guarded with Quantcast information as it apparently is with the existing server-based variety. In fairness to TBD, let me say that Quantcast’s estimates would appear to be drastically undercounting monthly visitors, and for all I know, Alexa might be doing the same—except we’re in the dark since TBD won’t give out the server-based stats (at least not to us nonadvertisers). Hey, TBD, do it! Lead the way through the numbers patch—with the openness missing from many and perhaps most news sites. How many visitors does TBD have according to Quantcast? I’ll save this little goody until the end of the current post.
–The main New York Times domain, nytimes.com, shows up in a Quantcast estimate—not server-based. Average monthly traffic comes in at “14.8M estimated U.S. people” over a five-month period. Will the Times Quantify its site, then share the data? Who knows?
–The New York Times Developers Network is Quantified and sharing the data—1.2K monthly U.S. people and 1.9K with overseas traffic included. Great. So why isn’t the regular Times site this helpful? Just an oversight? Perhaps the Times can explain.
–The Washington Post is Quantified but is hiding its traffic stats from the public at large. I’d love to know if the Post Quantified itself to squelch the estimate that Quantcast otherwise would make public without scripts running on the newspaper’s servers. Advertising prospects—all of them?—apparently can ask for the information by clicking on a link and providing email addresses; and I was about to respond to a related email confirmation request, but for legal and ethical reasons, I decided to avoid going further. Point is, the information should just be there, out in the open. The Post-related Quantcast page is the one from which I reproduced the message saying, “Traffic data has been hidden by the owner. Request access.”
–The Washington Examiner, Quantified, attracts 2.1M monthly visitors, all but about 100,000 in the States.
—The Washington Times is Quantified and draws 2.3M in the U.S. and another 200K or so outside.
–The ever-nosy Gawker gossip site is Quantified and is drawing 5.9M people, 4.2M in the States, while the related media network is attracting 17.5M domestically or more than the 14.8M domestic estimate for the main New York Times site (the network’s global count is an incredible 30.5M). No hypocrisy from Gawker when it comes to the Web numbers. Hey, nice going! Of course, disclosure is easier when you’re on a roll.
—Solomonscandals.com is Quantified, with public stats. Monthly traffic is a stinkin’ 1K (overwhelmingly in the States), down from a peak of around 1.9K. Perhaps I should be writing more about the adventures of Post legend Sally Quinn, son Quinn Bradlee and his fiancée, but, with apologies to Ms. Quinn and the happy couple, the hyperlocal scene is of far more interest to me right now. I’ve got the elitist view that traffic considerations alone shouldn’t influence news judgment. If/when The Solomon Scandals novel scores, the traffic will come anyway. TBD’s Jim Brady has some wise thoughts on the the need to strike a balance between numbers fixations and avoidance of useful measurement tools; also see a column from Washington Post media writer Howard Kurtz.
Meanwhile here’s one irony to contemplate. News organizations are eager to know all kinds of details about their readers; in return why the devil can’t they share with us their traffic numbers, whether or not we’re advertisers?
Wait. Here’s another twist, at least in TBD’s case, the one with the stat I promised. Because of small sampling sizes (Coumbia Journalism Review link), services like Alexa can undercount hyperlocal sites, just as Quantcast’s estimate has done through external measurement of TBD’s traffic. Based on the five-month trend, Quantcast says TBD is drawing just 5.7K U.S. people a month, surely a tiny fraction of the real number. Part of that severe undercounting might be due to the actual operational news site’s having gone live only on August 9. But I bet that sampling size would be a factor as well. Whatever the reason, the estimate-based graph above simply does not make sense.
If local and hyperlocal outfits like TBD will publicly share detailed, verified numbers from servers, they will be among the biggest winners, given the sample-size issue. But even giant sites like the Post’s will come out ahead, considering all the prejudices that some powerful trogs in the ad business have against the online media. Away with secretive traffic information and voodoo-style measurements! Get the real numbers out in the open—ideally with the means to share real-time information with Alexa-type services and even with the public directly. In the CJR piece, there’s mention of the possibility of a centralized measurement authority. I myself think that access to the numbers, with related technical standards and transparent methodology, would get us 75 percent of the way there even without Numbers Central.
For now, let’s consider a quote from Kate Downey, director of “audience analytics” at The Wall Street Journal and a subscriber to a bunch of measurement services. She told CJR: “People use whatever numbers look good that month. It gives publishers some flexibility. I think if everybody had the same numbers, we would hate that even more." Maybe. But with full access to the statistics and methodology, why not let advertisers call the shots, so that there would not be one number anyway—just more confidence in online news sites?
Update, 1:39: Or how about a mix of custom measurements and, for comparision’s sake, standard stats? Advertisers want to be able to help themselves and to be helped, not simply marketed to in the ways most convenient to publishers. With more open and trustworthy stats, advertisers will be better able to line up the best outlets for them (to everyone’s benefit, the media’s included). "Advertisers," of course, could include agencies and a variety of other kinds of people evaluating news sites for placements.
Good post on an under-reported question. Two points:
1. Looks like Quantcast defaults to actual “people,” which is a much smaller number obviously than visitors or pageviews.
2. I’d actually like to hide some parts of my Quantcast numbers, and I’ve seen others do it. I have no advertising, so its no one’s business but my own. But, the instructions on the QC site to do so are not accurate. Can you help?
Thanks, Chris. I dropped by Work, Wine and Wheels. You’ve got an interesting, well-crafted site with upscale demographics much like those of the Scandals blog; and if I were you, I wouldn’t worry about hiding a single stat in the case of WW&W. Not sure about your other sites. Are you worried about competitors analyzing you, or is it pure pride? I can’t help with the how-to info, but I suspect that Quantcast itself would love to hear from you so it could improve the documentation. If you follow through, let us know what happens. Thanks. David