Local searching on web-wide search engines can be disappointing. New moves by some major players may improve this. This first of a three-part series looks at local search listings that Overture is testing.
When I need a local product or service, I get low-tech but efficient. I pull the yellow pages off my bookshelf. Within a few minutes, I’ve found companies that can meet my needs.
Surely web-wide search engines ought to be just as efficient at finding local commercial information! Sadly, no. I’ve always found that searching the web for local commerce results is often a disappointing experience.
Fortunately, there are signs that local search is about to get better. Overture is testing a local search product, while Google is beta testing its new Search By Location service. In addition, long-time local search player CitySearch is expanding its efforts. These developments and others could mean that I’ll stop reaching for that printed yellow pages in the future — and so might you.
In this article, we’ll take a closer look at the problem with local searching in general, in terms of general purpose web search engines. Then we’ll examine Overture’s moves in the local space. Part 2 and Part 3 look at the work Google, CitySearch and others are doing.
Before looking forward, let’s see a few examples of how local commerce searching can be a poor experience on the web’s major search engines. We’ll start with a search for san francisco dentist on Google.
When I tried this recently, the number one ranked listing “San Francisco Dentist Directory” looked promising but turned out to have no listings at all. Instead, it was just a shell for listings that may or may not come in the future. The same was true for a top listed page from the “Internet Yellow Pages World Wide.” Another top listed page from a cosmetics site purported to show information about “San Francisco Cosmetic Dentist.” Instead, it merely provided links to a variety of non-San Francisco, non-dental cosmetic sites.
Make your search plural — san francisco dentists — and you’ll find more disappointing results. A page in the top ten from Zeezo.com promised in its title, “San Francisco Dentists-Dental: Dentists in San Francisco.” Yet, neither of the two listings on the page were for San Francisco dentists. It was a terrible listing for Google to top rank. Worse, this bad result gets rewarded by making money through Google’s own AdSense program.
Google also listed a page from dental-4.com promising “Cosmetic Dentists San Francisco California.” In reality, the page appeared to promote only one particular practice. The top results also had similar looking doorway-style template pages from the dental-2.com and dental-6.com domains, promoting two other practices.
What’s wrong here? Unlike with yellow pages, there’s no cost involved with getting listed in Google’s web search results. Consequently, there’s no natural barrier preventing anyone from trying to get a top ranking, regardless of whether they actually have content relevant for that term.
This issue is true with any type of search done on the web. People try to spam and game the search engines for all types of traffic. But local search is more susceptible to spam, I feel, because link analysis doesn’t work well to prevent bad content from winning top results.
There are fewer people creating links about the topic of “san francisco dentists,” which in turn means Google has to depend more on other, more old-fashioned and spam-susceptible means of ranking, such as the location and frequency of terms on the pages themselves.
I singled out Google, but it’s hardly the only search engine with below par results. I searched for san francisco dentist at Ask Jeeves. The same disappointing “San Francisco Dentist Directory” page that was number one at Google was also number one with Ask’s “Web Results” (which are powered by Ask Jeeves-owned Teoma). That page also got top ranked by AllTheWeb. Shifting over to the plural “san francisco dentists” found a page from the aforementioned dental-6.com site listed in the HotBot results from Inktomi while the Zeezo.com page also mentioned appeared at AllTheWeb and Ask Jeeves.
Also notable was a Sport Illustrated article that came up from eLibrary.com in the results at both Ask Jeeves and Inktomi. The article was a profile of a high-tech yo-yo expert who is also a dentist in San Francisco. Aside from not being relevant, the listing also stood out because it’s present through paid inclusion deals with both search engines. So, even adding a cost barrier is no guarantee of relevancy of local commerce search.
To be fair, mixed in among these disappointments are indeed some dentists actually based in San Francisco. However, the lists are not comprehensive. Nor do they deal with the issue of localities being near each other. For example, someone living in the Southern California city of Newport Beach might be also interested in dentists in nearby Huntington Beach and Costa Mesa. Unfortunately, there’s no easy way to automatically have these appear.
Mapping is another issue to be addressed. Want a hotel in San Francisco? Personally, I head over to Expedia. I know the travel booking site will give me results that really are hotels. I also get matches on a map, making it easy for me to discover what’s near a particular area of the city where I may want to stay. Do that on a regular search engine? No way. Well, at least not yet.
In September, I mentioned that Overture had unveiled a glimpse of its long-awaited local search listings on its Overture Research site, producing a flurry of articles that followed. Overture’s since shut down that preview, saying it was primarily being used to test some of its mapping technology.
Instead, the flagship for Overture’s local search product is now on Overture-owned AltaVista. A random sample of surfers going there, about 10 percent, have been tagged with cookies and now may encounter local search results when they do relevant queries, Overture says. Specifically, anyone who was presented with an option to do a local search during some early testing got a cookie.
For example, if you were in the sample group and doing a search for “san francisco dentists,” you would see a special “Local Sponsored Matches” section at the top of your results. Within this area would be links to some paid advertisers targeting that that phrase, plus you’d be able to click to see even more local results.
Choosing to see more brings up a familiar top ten list of local results. However, the listings are also mapped visually across the city of San Francisco. The map lets you interactively change the listings displayed. Zoom out, and Overture understands you want to see listings relevant to the San Francisco Bay Area, rather than just within the city itself. Click on the bottom of the map, and your listings move southward. All this happens without needing to enter any new city names, helpful especially if you don’t know the names of surrounding cities.
Users may also get invited to view local listings if they haven’t entered a location. For instance, a search for “flowers” might cause users in the sample group to be asked, “Would you like local results?,” with an option to enter a US city, state or zip code.
Who’s in these local listings? Not ordinary Overture advertisers. For example, you won’t find any of the listings in local search that come up for a regular search on Overture for san francisco dentist.
Instead, Overture has a separate database of listings that involves a small number of its US national advertisers taking part in a pilot program. Additional “backfill” results are also provided by yellow pages and data provider Acxiom.
Why not involve all the existing Overture listings? One reason is that the program is still being tested. However, a key issue is that those in the local program need to provide physical addresses to be associated with their listings. Without these, the listings cannot be mapped, a key part of the local product.
Take computer retailer CompUSA, one of the advertisers involved in the Overture pilot program. The company has hundreds of outlets across the United States. In order to show up for a query such as “compusa in san francisco,” Overture’s program needs to know the exact address of the store. With that address, Overture can then locate the store geographically on a map plus understand that perhaps it should also show up if someone zooms out from the map and wants to see all CompUSA stores in the entire San Francisco Bay Area.
So, Overture local listings will be a separate product that advertisers will need to purchase separately, just as some may currently choose to purchase country-specific products.
For example, those bidding on terms in the United States do not automatically appear for the same terms in the United Kingdom. Instead, they have to open a separate UK account and compete for terms against all others targeting the UK. The same is true for the new local product. Advertisers who want to be in it will purchase these listings separate from any existing keyword-linked campaigns they are running, and the pricing will be independent of that program, as well.
Naturally, Overture wants to make it easy for advertisers to take part in both programs. They company says it is developing tools to help its existing advertisers “add-on” to local, if it makes sense for them.
“It’s very important to us that we offer our advertisers a very simplified way of taking their bricks-and-mortar locations and getting them online,” said John Ellis, senior director of market strategy with Overture.
As said, “local” and “national” will be different products. This also means those thinking of going “local only” will need to carefully review where the local listings appear, as Overture partners begin to roll them out. They will not automatically show up in the same places as existing national listings.
For example, Yahoo currently carries the top four Overture listings for a particular phrase at the top of its results page in the Sponsor Results section with an additional two appearing in the More Sponsor Results section at the bottom.
If Yahoo adds Overture local listings, it might follow the style it already uses to alert users to yellow pages listings — a small link within the Inside Yahoo section of its results. Many users ignore these links and instead focus on what they perceive to be the main results. Given this, those going solely local might lose some visibility.
What exactly Yahoo — and other Overture partners — will do remains to be seen. It could be that local result implementations will be very compelling to advertisers. It remains too early to know. Overture also says that so far, the feedback from advertisers is that they’ll want to do both programs, not just one.
“It’s not really an either or. There is a lot of demand among our existing advertiser base is to buy both. For example, a national electronics chain may want to buy national Overture listings for their web site but have their [physical” store locations show up for our local product,” Ellis said.
Another issue is how to handle companies without physical storefronts. Take 1-800-Flowers as an example. The company can send flowers to any location, so technically is it local everywhere. But how do you map it? And how do you prevent virtual businesses like this from overwhelming the results when someone wants to see only real bricks-and-mortar locations?
Overture doesn’t yet have the answers to all of this.
“There’s a lot of interest in local, but I haven’t seen a lot of clarity on is what is ‘local.’ You bring up a great example of this concept of an ecommerce play that wants to deliver products and services to a specific location,” said Ellis, of the 1-800-Flowers scenario. “Destination targeting is one example, someone like 1-800-Flowers who wants to target geographically. Then there’s truly local, a physical distance that a consumer is willing to travel to acquire products and service. At the end of the day, it’s not strictly bricks-and-mortar, but we’re trying to create a system that lets advertisers drive offline transactions, to have a relationship with the consumer rather than have something shipped to their house, which is largely what online advertising is used for today.”
Another reason for this desire is because Overture already has relationships with many national advertisers like 1-800-Flowers, 1-800-Dentist or home improvement referral company ServiceMagic. What it lacks is relationships with many smaller, individual companies in specific localities that currently reach out to consumers through yellow pages. Local search is a way of bringing in new advertisers, not just making more off of existing ones.
Overture expects its yet-to-be-named local product to open to any advertiser in the next several months. When that happens, there will obviously be some queries lacking listings. To avoid disappointing searchers, Overture will partner with a yellow pages company to provide “backfill” answers for when it has no paid local listings of its own. A contract with a firm for this has been finalized, but Overture’s not yet releasing the name.
By the way, Overture is not using another backfill option available, the crawler-based technology of Lasoo. Lasoo is a company I wrote about back in mid-2001 that allowed you to view web search results presented in map format.
Overture acquired Lasoo in the latter half of last year. It’s using the Lasoo mapping technology, rather than the crawling technology, as part of its local product. Also, while Overture acquired Lasoo, it did not acquire the Lasoo.com domain. So that site is not owned or associated with Overture.
As for how and when local search goes live with partners — and even who those partners will be — that’s all still being negotiated.
“We’re in active discussions with all existing and some potentially new partners. There’s a lot of interest and a lot of demand. We’re not disclosing specific partners that have yet signed up,” Ellis said.
It’s important to remember that the AltaVista implementation does not reflect what other partners may do. For example, a partner might decide to show all local results rather than just a few, when there is confidence that the query is local in nature.