I took some vacation last week and missed Steve Lohr's email asking me for comment on search optimization. Looking over the resulting New York Times story, "This boring headline is written for Google," a couple of things come to mind:
- We don't want/need Google driving traffic to everything.
- Those boring headlines were a priority long before Google, for reasons that have nothing to do with search (and everything to do with browse).
On the first point: Google can be like a fire hose. I have some recent experience with fire hoses, having drained a swimming pool last week. They're hard to point, and you don't necessarily want to get in their path.
Newspaper sites have very specific audience targets: local, local, local, and people who are about to become local. Newspapers are local businesses built on the proposition that they can deliver local audiences that are relevant to local advertisers. While there may be some ego satisfaction in gross traffic figures -- unique users and pageviews -- it's the net local audience that counts.
All traffic costs money -- content server expenses, and ad server fees. Nonlocal traffic costs just as much as local to serve, and generates little in return. When local ads are delivered to nonlocal users, clickthrough -- one of the measures examined closely by paying advertisers -- drops substantially.
A newspaper with a good registration system can distinguish between local and nonlocal users and, with suitable ad server magic, can conditionally serve nonlocal ads to nonlocal users -- but there's not much incentive to chase those nonlocal users. The newspaper industry lacks an effective national online sales system. There's some minor money in plugging into junk-inventory networks, but much of the ads that come through those channels are the kind that drive users away.
So the last thing a local newspaper website needs is a global search engine indexing its nonlocal news content. Years ago I had the techies block Google from spidering AP Online data on my employer's web servers.
There's good reason, however, to have Google (and others) index genuinely local content, because some substantial percentage of people who arrive at the website through search links will be genuinely valuable. (It's not guaranteed, though. I noticed the other day that one of the most highly ranked pages on Bluffton Today's website is a user blog item about getting cat pee out of the carpet.)
The most important thing to prioritize, of course, is the "revenue verticals" -- local jobs, homes and car databases. Looking across the industry, it's a sorry mess. Often a real estate page is put together by an advertising department that doesn't understand how search engines work (or, apparently, users). Many sites are laden with huge graphics, which are ignored by search engines and generally hinder users. What's missing is the attention to detail in the language -- expressed in HTML text -- on the page.
Search engines index content. If the pages don't have text content that accurately describes the resource, no amount of monkey business with titles and meta tags will fix the problem.
Now, about those boring headlines. It's all about the browse, not the search.
I learned a long time ago -- and learned the hard way -- that a dull, boring, but informative and accurate label beats the socks off a cute pun when it comes to driving clickthrough in an interactive context.
It should be obvious that the online experience is fundamentally different from the print experience. Despite the efforts of jump-crazy news editors, a newspaper page still has a significant amount of reading matter embedded directly in the user's line of sight. A cute headline works in connection with a lead and (hopefully before the jump) a nut graf.
But on the Internet those headlines have to carry the full burden of communicating enough information about the item to enable a user to make a "purchase decision" -- clicking through to the content. So they'd better get right to the point.