In Seattle, the Globe creaks noisily on its axis.
In the PI newsroom, remaining staffers are awaiting a clemency call they know will never come, or as one told me today, "I'm one of those semi-miserable souls who is waiting for the word from Hearst on when the print product dies." The current death penalty appointment date: Most are betting on next Wednesday, March 18, the original date provided in the federal WARN notice that Hearst filed.
Watchers everywhere are left with the sport of pondering what the heck Hearst is thinking (Mark Potts' wonderment) and imagining conversations between Hearst execs (great speculative business modeling from Nieman Journalism Lab's Martin Langeveld). That modeling is key, of course, to the next stage, the new online battle to come. I provided a view of the current head-to-head battle between the PI and the Seattle Times, in this post.
Dave Boardman, executive editor of the Times, takes issues with the numbers in that post, and I'm publishing his thoughtful response, in full below. He also points out that I've over-estimated the fast-dwindling Seattle Times newsroom staff, which is in fact now about 210. My 260 number goes all the way back .... to November, according to this Times story, another sign of how quickly we're seeing the deceleration of the business. Sorry for that error and for all those staffers who lost those jobs.
Quickly, to the traffic number controversy. Everyone in the trade knows that traffic accounting is a quagmire. There's Nielsen, Comscore, Alexa, Quantcast, Compete, and newspaper's industry's longtime standard, Scarborough. Check Seattle Times and Seattle PI traffic on all of them, and you'll see how wildly the numbers vary.
Then, there's the internal site counts, which I often hear about in talking with site managers. They have their biases as well; any metrics system does. Of course, those are unavailable to the general public.
So, in this case, I used Nielsen, with the notion of an apples-to-apples comparison providing some reasonable basis for appraisal. Nielsen is after all what the Newspaper Association of America --- the industry's overarching trade group -- uses.
Certainly, within Nielsen, as well, there's much room for interpretation. How much of that traffic is local/national; how much is SEO (search engine optimization) responsible for traffic, etc?
Even with Times' own numbers, the two sites still appear fairly close in most metrics, with the major disagreement coming on the basis of what the Times considers it stronger "local" penetration. The impact on advertisers is, of course, another story again, as the dirty, little secret of many "local" news websites -- for more than a decade -- has been how much of their web traffic is non-local.
Boardman's comments about the loss of the PI from the current network sell are sure to be taken as fighting words by the (presumed) new competition.
The new online PI will not use the Times to sell any of its online advertising, I have confirmed. So let the smack-talking, ad-selling and c-o-m-p-e-t-i-t-i-o-n begin, finally. Enough anticipation.
All of the traffic issues deserve fuller exploration and -- someday -- greater resolution and transparency. (In this case, we're not even sure, how much the PI has seen the numbers cited below.) For now, though, let me let Dave Boardman have his say:
I'm even more disappointed that you didn't even make a phone call to us to check your facts. Among the errors is that we have 260 staffers, which overstates reality by almost 20 percent.
There is a fundamental problem with using Nielsen's numbers as a starting part of any analysis. Nielsen is wildly and consistently inaccurate when used to measure regional sites such as seattletimes.com.
Nielsen depends on monitoring software installed on individual computers spread across the United States. Then, the numbers Nielsen reports every month are extrapolated or “upsampled” from that relatively small group of people. The sample size may be adequate when it is used to count sites that have a fairly uniform national reach, such as nytimes.com or usatoday.com. But when a site is primarily local, the sample size is tiny. That explains the wild swings in the monthly numbers and the inaccuracy in most of them. It's not big stories that create the differences, it's a methodology that doesn't work when applied to anything but the largest national sites. This is even more of an issue when a site is specifically targeted at local and regional users, as seattletimes.com clearly is.
Compounding that problem, Nielsen has admitted that because most companies won't allow the installation of the monitoring software on corporate computers, very little monitoring is possible during working hours, the time when most online news consumption occurs.
The result is massively underreported results, especially in unique visitors. Our Omniture readings show us at about 5 million unique visitors a month, as opposed to Nielsen’s reading of less than 2 million. (Nielsen also vastly underreports pi.com’s uniques, as well.)
In your bullet point about page views, you admit that the P-I advantage looks odd. It looks odd because those numbers are completely wrong. Our page views as measured by Omniture were 45 million for the month, not 11 million. Again, because we administer the servers for both sites, we know the actual traffic for the two sites is neck-and-neck and the lead varies month to month. The actual data show nothing like the variations Nielsen reports.
According to Scarborough market data, for December 2008 in our Designated Market Area — the local audience most watched by advertisers — p-i.com had an audience of 347,000 and seattletimes.com had 592,400. That ratio has been consistent month to month, and the totals have steadily grown since the launch of both sites. And notably, our market data shows us most of those pi.com users also use seattletimes.com, so the loss of pi.com from our online advertising network will be of little consequence.
Finally, as I am sure you recognize but don’t acknowledge here, not all hits are created equal from a business standpoint. The P-I relies heavily on photo galleries of events outside the Northwest – New York and Paris runway shows, beauty pageants, swimsuit models – that produce tens of thousands of page views but offer little revenue or lasting relevance. In fact, had you spent any quality time looking at the two sites, you would have little doubt that The Times, not the P-I, has more local news content – and more value – for users."