If I ran a bagpipe band, I'd call our first album The Scotch Tapes.
Three (plus) things about local search
As I mentioned, I was a panelist Tuesday in a session at the Newspaper Association of America marketing conference, titled "Searching for the Holy Grail in Local Search."
The panel was a blast. Greg Sterling rules as a moderator (you can also read his take on the panel).
We panelists were asked to come armed with two things: our definitions of "local search" and a list of three takeaway points for newspapers wanting to engage in local search -- things we have learned while working with local search products.
I didn't lug these bullet points to Las Vegas in the form of handouts (I refuse to kill trees needlessly on behalf of digital media), so the least I can do is provide them in Blog-O-Matic format.
My definition of local search: Rather than say it's products that enable searching of all forms of locally relevant and interesting data, I focus my definition on the types of local information where the greatest concentration of business opportunity seems to lie:
- Business directories, either general interactive Yellow Pages, or vertical niche directories such as the Tri-State Home Show site launched by Scripps' Evansville Courier and Press.
- Local government and public agency directories, including schools, parks and other community institutions.
- Local guide products, including arts and entertainment events, restaurants, community sports, civic and social organizations.
My three takeaway points: Not necessarily in order of importance ...
- Newspapers' biggest competitive advantage in local search products is human editing of data by people who actually live in the markets served by the products. No directory data source, filtering mechanism or search algorithm can substitute for that.
- Human editing is hard. It will almost certainly cost more, be harder to recruit for, and expand the roles of your existing news people beyond their comfort. Do not assume that purchased data alone, or self-service tools that allow business owners or event sponsors to post and edit their own stuff, will cover the base for you. Local search businesses will fail if the data management is put on autopilot.
- Don't try to build what you could buy, license or form business relationships to share. Most of the required functionality and componentry for local search products is done already, by a choice of service providers in many cases. Why pay for and try to support a one-off solution if you can't point to a clear differentiator in the resulting software? Cool user interface tricks are fleeting competitive advantages at best. And the command-and-control mentality that leads to a "stick-built" development strategy usually also leads to software that arrives at Version 1.0 and never gets past it.
I'm happy to expand on these points if you give me a shout or post your questions/reactions in the comments. I'm off to board my plane home, happily. I've had fun in Vegas but five nights is too long to be here.
Update (7:10 a.m. 2/2/07, and yes, I'm back home): Peter Krasilovsky and Mike Boland offer their interpretations of the local search session.
"Once a machine is
"Once a machine is programmed well enough to do a good job, it does so the same way over and over at a low cost."
It's the "do a good job" part where humans still play a key role, Stephen. I would argue we're a long way from Web databases knowing without prompt when new restaurants open, old ones close, owners change or bands cancel a gig. (I'm not sure I ever want machines to get that smart, but that's just me.)
Humans still have to report those occurrences to machines somewhere in the value chain, and then the Web databases that we use to make business and event data into local search services have to know where pick up the data.
In some cases, it takes an intermediary. Owners whose businesses fail, for example, have almost zero motivation to notify local search engines that the listing is obsolete. The phone company knows a number has been taken out of service, but not always why. An enterprising person, however, who drives by and sees the out-of-business evidence can pass that update along if given tools to do so.
"No directory data source,
"No directory data source, filtering mechanism or search algorithm can substitute for [human editing of data by people who actually live in the markets]".
Oh, how I wish this would be true forever but it won't be.
Don't forget that once a machine is programmed well enough to do a good job, it does so the same way over and over at a low cost. Humans do things depending on their various skill levels and even mood, etc. They also cost a lot more per operation than a machine.