Marty Himmelstein’s previous post “Why We Don’t Have Good Local Business Content?” sparked considerable discussion and debate. I’m letting Marty respond to the comments (in two separate posts).
The first about the role of SEO in local search is below. The second one, according to Marty, “will address some of the questions on the role of the community in local search and gathering content directly from businesses.”
I have not edited the text.
Several weeks ago I wrote a guest post, “Why We Don’t Have Good Business Content.” The post elicited a number of thoughtful responses, for which I am grateful. Instead of replying with additional comments, Greg was gracious enough to give me the floor again for a couple of follow-up posts. This one covers the role of search engine optimization in local search.
I remarked that when local search is working as it should, the role of SEO will be diminished. Andrew Shotland responded to my statement that in local search “being found trumps search engine optimization” with a pithy comment to the effect that being found requires SEO. I agree with Andrew with the understanding that he is describing the current situation, and my intent was to describe why the current situation is broken. Andrew’s observation will continue to be true until we have an accurate layer of content about local businesses, and, more generally, local places.
Web search and local search start from two very different places. In web search, we have essentially infinite content, and the process of evaluating and ordering that content are inseparable. Local search doesn’t deal with infinite content. It doesn’t even deal with a lot of content. The necessary underpinning of local search is not evaluation, but fidelity to physical reality.
With web search there is no purely objective measure of goodness for a particular query. Therefore, search engine designers devise algorithms to approximate the best pages from a sea of content. These algorithms employ subjective measures of worth that involve tradeoffs amongst competing goals. SEO, when used appropriately, can be seen as a way to help the search algorithms do a better job of ranking pages fairly. Sometimes SEO might be justified on the grounds that the unaided algorithms just get it wrong. On the other hand, when used injudiciously, SEO has the effect of subverting the algorithms, which, while imperfect, are usually pretty good.
Local search is simpler. The criteria for local search are factual and objective and it is therefore possible to get pretty close to unbiased results. Does the place exist where it says it does? Does the place do what it says it does? There aren’t twenty million plumbers in my local area. (The number of results returned by a Google search for plumbers.) There aren’t even dozens where I live. In most places, the relevant listings can be displayed on a map, on the first page of results.
But in NYC there are well more than a dozen plumbers, and whether one is two-tenths of a mile or two miles away hardly matters. In the strictest sense, then, Miriam Ellis’s observation that “any list-type listing means competition … even if it’s only alphabetical competition” is true. In a larger sense, however, a comparison between the minimalist ordering sometimes required for local search and the sophisticated analysis needed to prune 20 million results demonstrates the differences between the two types of search. No matter how clever Google programmers are, they can’t design their algorithms to know who the good plumbers are. Or where the best sushi bars are. In any case, they can’t do as good a job as people can. In web search, because of the overwhelming amount of content, we need the core search algorithms to do as much as possible. In local search we want the core algorithms to do only as much as necessary. The quality of businesses is of course important, but the way we can discern quality is by associating user-feedback systems (and other third party sources of content) with the core data. In the evolution of the web, we’ve come to understand that sometimes there is no substitute for human computation. Local search is exhibit one.
Assume for a moment we had a reliable and complete stratum of local search data, a database of record for local search. This stratum would contain, as nearly as possible, a core of factual information for each business. This layer of content would be value-neutral. Its main purpose would be to maintain an accurate model of businesses and places in the real world. When new places open or close, these facts would be reflected in the database as quickly as possible.
Gathering a layer of factual content might be hard, or it might not, but gathering it is a distinctly different task than making value judgments about it. If this layer existed, review sites would have access to it, and they would be free to enhance and order it in any way they like. Different sites could order it differently, or people could request different orderings. One person might want only handicap accessible stores, or stores with wi-fi, or Goth friendly places, or whatever. In all cases, though, the underlying content is the same.
Another reason the role of SEO should be minimized in local search relates to trust. In local search, if one business is optimized, it is at the expense of another. Such optimization might have no intrinsic relationship to the actual worth of the two businesses. This apparent arbitrariness serves to erode the confidence of the two most important local search constituencies: consumers and small and medium businesses. The best carpenters in my area don’t advertise at all. I still want to be able to find them. Most of the time, when people are looking for a carpenter online they will want a recommendation. But sometimes they won’t: they will want the carpenter they used last year, or the one their neighbor mentioned. And, as I said earlier, when they want an ordering, it will be one supplied by other homeowners, not an algorithm that can be ‘optimized’ by the initiated.
Marty Himmelstein is the principal of Long Hill Consulting, which he founded in 1989. Marty’s interests include databases, Internet search, and web-based information systems.
For the last eleven years, Marty has been active in location-based searching on the web, a field often called Local Search. Marty was an early member of the Vicinity engineering team. Vicinity was a premium provider of Internet Yellow Pages (Vicinity provided Yahoo!s IYP service from 1996-8), business locators, and mapping and geocoding services.