The Web metrics jungle: TBD hyperlocal site beats WUSA and Washington Examiner in ONE local Web measurement derby

Posted by’s hyperlocal site is drawing more local Web traffic than WUSA-TV and the Washington Examiner and may close in on the Washington Times and Fox’s D.C. outlet—-if you go by area Web statistics from a major measurement service, Experian Hitwise .

But in local audience size, TBD is a long way from threatening the online supremacy of the Washington Post, even allowing for the upstart Web site’s newness. In my opinion TBD, although apparently growing, could come closer if the talented people there paid more attention to the suburban audience and otherwise refocused. Significantly, runs stories from WJLA-TV, Channel 7, and now redirects visitors to the startup site. Add in the free promos that TBD’s Web site is enjoying from the cable channel with the same name, and you wonder why TBD couldn’t fare much better than it is now.

imageExperian Hitwise kindly shared the D.C.-area numbers with the Solomon Scandals site after running across my Tweets about Web measurements. Of course, the numbers may or may not agree with those from other sources, which have their own samplings and their own methodologies. The D.C. area is a good example of the confusion now reigning in Web metrics, one reason why advertisers haven’t poured as much money into the Web as they should.

Here’s “Most Popular Web sites in Media” for the Washington market, with the geography defined by both Experian Hitwise and its rivals at Nielsen. You can better see the source of the information by clicking here or double-clicking on the above image to the left.

image 1. The Washington Post enjoyed 71.08 percent of the market for the “four rolling weeks ending September 18, 2010.” “Market” in this case reflects the number of visits, not different visitors. One person could generate multiple visits. Also remember that this is a survey of the Washington market. Nationally and internationally, I suspect, that differences would be even greater between the Post’s Web site and most of the others mentioned here. But these local numbers will still mean plenty to advertisers focused on the D.C. area.

2. NBC Washington, aka WRC, Channel 4, came in far behind the Post at 6.94 percent, despite the TV site’s strong interest in at least certain categories of local news, such as politics and the social circuit.

3. The Washington Times commanded 5.85 percent of the local Hitwise-defined scene, although national and international percentages might be higher, reflecting the paper’s current priorities, which have dramatically reduced local coverage.

4. MyFox of Washington, D.C., aka WTTG, Channel 5, drew 5.09 percent.

5. TBD came in at 4.46 percent, not that far behind the local Fox outlet and the Washington Times, if you consider that it’s just starting up and will probably grow numbers. Of course, that’s still a long way from the Washington Post, as noted; and once again, remember that the survey is focused on sample Washington-area surfers. And keep in mind, too, the WJLA-originated resources (see Item #8) that TBD can coast on. The good news is that TBD has lots and lots of potential in areas such as suburban coverage, education coverage and genuine neighborhood-level reporting.

6. Gannett’s WUSA, aka Channel 9, measured 4.29 percent—a less-than-stellar showing if you consider the station’s past strength in news coverage. Shame on Gannett.

7. The Washington Examiner is at 2.14 percent, a little unfair, considering the breadth of the local news coverage, given the paper’s limited resources (will the owners acquire a TV station to—among other things—promote the Examiner’s coverage?). This excludes traffic from Washington-related pages in the national network. Speaking of local media outfits with limited resources, I hope that the Washington City Paper’s Web site can be included in future surveys from Experian Hitwise, even though I doubt that the percentage will be large. Another possibility for inclusion would be the total for D.C.-area sites from, AOL’s hyperlocal startup.

8. Allbritton Commuications’ VHF station, WJLA, aka Channel 7, is last at a miserable .16 percent. But so what? Remember, now redirects surfers to And as noted, the new hyperlocal site shares Channel 7’s local reporting—in addition to the work of the dozen or so TBD reporters, plus the resources of the former NewsChannel 8, which TBD’s TV side used to be.

Now—about methodology. Experian Hitwise did not provide its sample size, but whatever it is, I suspect it is larger than the ones that used in coming up with the statistics that I quoted earlier. Yes, the Alexa-based headlines here about the TBD local experiment have not been as optimistic as the current Hitwise-based one. But you know what? Other stats I dig up might contradict those of both Alexa and Experian.

All kinds of questions arise. I myself like the idea of advertisers and the world at large having instant access to a variety of statistics, rather than just certain “standards.” In particular, I want to see server-based measurements, which would probably be more precise than sample-based ones and which would be especially useful in measuring the audience for little hyperlocal Web sites (in fact the Quantcast Web measurement service does provide server-related numbers, and several major D.C. area sites already participate, including those from Experian second-placer WRC, as well as the Washington Times and the Washington Examiner).

I’d even like see ZIP code-based measures originating from server-based technology. Since Internet addresses reflect geography down to the neighborhood level, this should be possible—if not now, then in the very near future if advertisers are smart enough to insist on this.

Better metrics would also help news organizations, more precisely rewarding performance. Even at the metropolitan level, sample-based numbers won’t necessarily give you the full story because of the sample size factor. That said, I’m super-grateful to Experian Hitwise for sharing its data, which I suspect is far, far superior to Alexa’s for the D.C. area.

Meanwhile, within the limits of time, I will keep trying to:

–Get server-based statistics out of TBD, which has boasted of its openness and registered with the Quantcast, which, as noted, uses server-based information. Quantcast participants can choose to share the numbers with the public. TBD Community Engagement Director Steve Buttry is to get back to me with an answer as to whether TBD will go public with its forthcoming Quantcast stats. Currently Quantcast runs estimates for TBD, but even they will vanish if TBD chooses to suppress the numbers. I suspect Steve himself would release the information immediately if he had sole control. Why should media organizations, with all their blat-blat about openness, be like the Pentagon, where the default so often is secrecy? Just what could be a more basic issue than audience size? The goodies should be available to the entire world online, not just advertisers or media reporters. Open stats—and open methodology, which most sampling services almost surely lack—are an excellent way for the online media to build credibility. The methodology could take into consideration the fact that some of the visits to servers are from other computers, such as search-engine spiders, rather than human. Some numbers geeks have raised the possibility of “an open cookie standard.” Cookies identify your Firefox, Google Chrome or Internet Explorer browser to the Web sites you’re visiting.

–See how easy it is to obtain Nielsen information for the Washington Post. A Post spokesperson tells me that the information is public, and maybe it is if an established media organization requests it. But so far, to my knowledge, Nielsen’s PR people have not returned my call for the numbers. I’d love just to be able to pull the statistics off the Web. But I haven’t found the Post’s Nielsen numbers online at the measurement company’s site—am I missing something? Maybe. As for Quantcast, the Post participates in the service but deliberately hides the numbers from the public, preferring to standardize on Nielsen. And you know how far I’ve gotten on that. I’ll also be curious if Nielsen has server-based numbers for the Post. Come on, Post people. Given the problems that other newspapers have had with the sample-based approach, you’ll be much better off with numbers from the servers—ideally verifiable through the Audit Bureau of Circulations. And get the Quantcast numbers up there, too.

image The actual numbers, not just the percentages of the market, are of greatest interest to me. Percentages are nice to have. But if I were an advertiser throwing out X number of dollars—whether I was a neighbor restaurant owner or GM’s ad agency looking for precise targeting—I’d care most about the number of eyeballs I was grabbing per dollar. And with goodies such as ZIP codes and even locations within neighborhoods available, I’d find the numbers even more useful and be more likely to spend on local and hyperlocal media. Private stats aren’t as good as public stats, with the ability for the entire universe to discuss and debate the methodology.

While the advertising business still has its share of Neanderthals, it is entirely rational for these Luds to distrust the new media if Web sites do not give the full story. Why haven’t some members of the new media cared more? Could it be that some important Web sites are owned by old media companies that would rather that the Web not be too attractive a place to advertise? I don’t know. No accusations! But if people like Washington Post publisher Katharine Weymouth want to live up to their platform-neutral philosophies, then it is high time to get those server-based stats online ASAP through Quantcast and other means. The news side will probably cheer this move—especially, I’d hope, business geeks like Post Executive Editor Marcus Brauchli. I’ve read of the ”Brauchli doctrine,” the belief that the press wastes too many precious minutes explaining itself and its actions. But ideally traffic counts would be another matter. Hey, just automate Web reporting or at least open up the Quantcast statistics, the way NBC Washington, the Washington Times and the Washington Examiner have done.

The Quantcast situation, with media mentioned in the order I used earlier: (statistics hidden from the public), WRC/NBC (unique monthly visitor count of 893.4K national, 929.3K global), Washington Times (2.2M national, 2.5 million global, U.S. included—in keeping with my earlier observation that the Times would come across better if you looked beyond the D.C. area), WTTG/Fox (not “Quantified,” although Quantcast does offer an estimate of 294.4K U.S. visitors, without a global number given), TBD (5.7K estimated American visitors, although I think this number is wildly inaccurate on the low side—one more argument for a server-based approach included), WUSA (173.3K estimated U.S. visitors, with Quantification in progress), Washington Examiner (2M national, 2.1M global) and WJLA (189K estimated U.S. visitors—an obsolete figure since since Allbritton has folded the site into TBD, but maybe a better stat for the new hyperlocal site than that crazy 5.7K count).

Related: 1 in 4 local firms plan newspaper ad cut: poll, a post from Alan Mutter in his blog, Reflections of a Newsosaur—the poll was national. The good news is that, according to Mutter, "71% of respondents to the ITZBelden/API survey said they are confident that newspaper ad reps could help them make the most of their marketing dollars on the web, social networks and in mobile media." In fact, spending on web, social and mobile media is on the way up (expected to average 13 percent of ad dollars in 2010—still far, far too low). All the more reason to encourage this trend through more trustworthy and detailed reporting of Web traffic! While most newspapers would be foolish to drop print editions immediately, the Web is where the future is. Measurement practices—and technology if necessary—need to catch up in the D.C. area and elsewhere.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.