Panel transcript from the Digital Directories & Interactive Local Media Summit, October 22, 2003, Alexandria, VA.
Gerry Campbell, Executive Director, Search and Navigation, America Online
Paul Levine, General Manager, Yahoo Get Local
Tony Mamone, VP Listing Services, LookSmart
Michael Oschmann, Managing Director, Mueller Medien
Tom Wilde, GM Search Services, Terra Lycos
Bobbi Loy-Luster & Greg Sterling - The Kelsey Group
Moderator: Could you define what local search is exactly and why is it so hot?
Gerry Campbell, AOL: "Local search is using all product capabilities we have available so that users can find what they want on a local basis. It's getting all of the pieces to come together to satisfy consumer."
Tom Wilde, Terra Lycos: Defining a "local search" is a bit of a challenge. There isn't a clear definition if "local" means a user wants results in a particular town, metro area or 50 miles from a location. You need intelligence applied to certain categories before you can determine if the search is local.
Moderator: What percent of searches are local in nature, and how do you determine that?
Paul Levine, Yahoo: From the main search bar, there are two categories of "local search". There are explicit local searches where the user has a locality (city, area, zip, etc.) specified in the query. Then there are implicitly local queries where the type of search terms only make sense given localized results. Examples of these would be searches for a dry cleaner or auto repair where ultimately the right response for that query has a local flavor. Chances are that somebody isn't looking for the history of auto body repair, they're actually looking for a local commercial business to satisfy that query.
Explicit local searches range around 5% of all searches. Implicit searches are much larger in the range of 15-30% and that's kind of industry wide. It depends a bit on which search engine your using. Within Yahoo YP, just about 100% of the queries are local.
Tony Mamone, LookSmart: When we first started attending the Kelsey conferences about a year and a half ago, we used census data to look for terms in a query string to determine how many searches were local. We concluded that 5-6% of searches were considered local because they had a geographical qualifier. We redid the analysis this spring and found that 8-12% of searches were local in nature. I think the increasing local search is a factor of users becoming more sophisticated in using local search terms and that we're getting better at delivering local search results. It's a little bit of a chicken and egg because if the results you get are poor, you tend not to use that service again.
Moderator: Paul, have you also seen an increase in local search?
Paul Levine, Yahoo: Yeah, we've seen a slight increase in local search terms but we don't track it on a regular basis. I absolutely agree with the point that the more you reward users with local search data, the more local search you'll get. We've been integrating Yahoo YP into search results to boost local search results.
Gerry Campbell, AOL: There is a gray area of search that has no explicit local search term. We've been focused on supplementing local search results by integrating AOL's IYP content into searches that are identified as having local intent. Interestingly click through rates increase as we provide more local data. This indicates that there is a pent up demand for local search results.
Currently, local searches range between 15-30% of all web search on AOL, but with the right technology delivering higher quality results, the local search will increase. Our experience today is directly parallel to what they've just said.
Michael Oschmann, Mueller Medien: We come for the other side where our IYP search was all local or regional in nature and we avoided any national search. Our users complained, so we've added national search to our products. Our local searches comprise 80-85% of all searches, and 15-20% are national search.
Moderator: This seems like a classic chicken and egg situation. Until you have robust local content, you can't attract local searchers. But without local searchers, how do you build the directory?
Tom Wilde, Terra Lycos: Changing user behavior is a difficult thing. Users don't use local search products with anywhere near the volume the do web search. Users come to a search site expecting to find what they're looking for with only one or two words, and they don't add these modifiers naturally. Getting people to tell you their location is a challenge. Without local qualifiers, search is disappointing. The reality is that there is not a meaningful critical mass of leads being generated by local web searches, so advertisers aren't willing to put the effort into advertising because use isn't there.
Gerry Campbell, AOL: There needs to be a payoff to change consumer behavior. The payoff is not about hype, it's about absolute daily utility for the user. The reality is that getting a critical mass of advertisers is very hard and that's why YP has the essential beginning. YP lists every business regardless if the business is an advertiser. You need enough people to build up the expectation that local searches will be resolved successfully. With demand increasing, we then have a value story to take to advertisers. The tipping point happens when consumers expect results.
Paul Levine, Yahoo: Some localized version of web search is not the answer to resolving local search queries. The IYP products that are already online answer many of the issues. When Yahoo thinks of local search, we don't assume that the main search box is necessarily the right place to start. The richness and functionality in the IYP may become an offshoot of the main search box. Local search exists today although the results are not perfect in web search or in IYP. We're moving toward a hybrid between IYP and better web search to provide better local results.
Michael Oschmann, Mueller Medien: I'm glad that the search engines don't have the answer, but I'm worried that they are going to resolve the hybrid problem. The problem is that there are too few leads at too little prices going to the local advertiser today. I think that the critical thing is to develop good local products for sales force to sell. We need to have high value products that will sell at high enough prices to justify the sales call and maintain high sales commissions. The problem will be to convince local businesses to pay reasonable prices so that we can afford to have the sales force call on them.
Moderator: What are the biggest obstacles to effective search localization from a user standpoint?
Paul Levine, Yahoo: The fundamental obstacle is that 80 - 90% of local businesses don't have a web presence. Trying to localize web search when the content isn't there requires a hybrid of IYP and web search. Beyond the technology piece, user education is necessary to change people's ingrained habits. Users need to be able to move form the structured data in the IYP to the unstructured data results in web search. The challenge is helping people understand what information resources are available.
We've found that consumers love the integration of IYP results into the web search. It's been very well received. More and more people are using that web search box to navigate into Yahoo YP. We've gotten very good feedback from users especially those who didn't even know we offered Yahoo YP. We're in the early stages and haven't done a great job of presenting depth of data in IYP. Users want more and more information.
Gerry Campbell, AOL: I would echo that completely. People love getting YP results in web search. It bridges that gap between structured and unstructured data. We have to focus on giving the users what they want. Looking forward, the challenge is to constantly stay in the users shoes to understand what the user expects. We need to know when they're searching if the results they see are commercial vs. non-commercial and informational vs. non-informational, and web oriented or local.
We're finding that people using MapQuest and Citysearch are looking for more general topics than they were in the past.
Tony Mamone, LookSmart: Getting data from local businesses is difficult. It's a challenge to gather keyword lists and other relevant data and then help users find them. It's a big leap for local merchants who don't understand web search. It's also hard for the sales rep. Need to get better at collecting data.
Michael Oschmann, Mueller Medien: I believe that it's difficult understanding what the users really want. In the past, as a traditional Yellow Pages publisher, we've gotten used to being distributed into everyone's house, with virtually no competition. It's very important to listen to what the user wants. The IYP is totally different than it was a few years ago. It's good news for traditional print YP because we find that people appreciate the structured headings.
Moderator: There are perhaps 250,000 paid search advertisers on Google and Overture with substantial overlap and there are roughly 10 million small to medium enterprises (SMEs) in the US. How do you tap into the rest of the market?
Tom Wilde, Terra Lycos: We call it the last mile problem. Without applying sales resources, it didn't take off. You have to remember that Google, Overture and Looksmart all have sales forces. The idea of self-provisioning can only go so far. I'm skeptical that self-serve can reach 10 million users. In Spain our sister company has 2,000 print YP sales people which is 10 times the sales force of all the IYP companies combined. I'll make a prediction that maybe Yahoo or Google will buy a 2nd tier YP publisher to get at this market.
Paul Levine, Yahoo: Particularly for the smaller enterprises it must be sold instead of simply bought. The rubber meets road when a sales rep walks into a SME. That sales call overcomes the operational challenges of buying online media. Yahoo's partnerships with BellSouth & SBC have proven to be extremely productive in positioning Yahoo's inventory with their IYP inventory to reach the SME. We think those types of relationships will be really effective in selling the local search products down the road.
Gerry Campbell, AOL: There's one more element. The sales force faces a hurdle as well as an opportunity with cost per impression (CPM) or placement in the book vs. cost per click (CPC) sales structures.
CPC is difficult to manage even for the technically savvy. Beyond the early adopters, you're trying to sell a very technical ad model to a very non-technical audience. I think you need some models that are more traditional where you buy an ad and it's there for a month or 6 months instead of expecting a small business advertiser to manage keyword bids.
Moderator: Realistically, of the 10 million figure, what's the real potential of this market?
Gerry Campbell, AOL: Over the next several years, several million SME businesses will be looking at this opportunity. Primarily, we need to prove ROI. Must be expressed in simple terms the SMEs understand. We need to reach out to the business. It takes time to get used to this model working.
Tony Mamone, LookSmart: The acceptance rate depends whether the new evolution of ad products is a separate upsell or if they are part of a bundle. That will determine if it's a proactive decision the SMEs need to make or if it is a ride-along. The fastest way to gain penetration is if it's bundled with something that they are already used to. If it's separate, adoption will be much slower. It's critical to bundle so that we get up to speed much faster. That gets you to prove your model and attracts users sooner.
Tom Wilde, Terra Lycos: I think you need a bundle. Requiring discreet decisions will take much too long. A good bundling strategy gets this to take off and then you can sell additional services, enhancements, mapping, etc. There's a limited amount of work required on the user's part.
Moderator: Michael, as a traditional print YP publisher, how do you react to this?
Michael Oschmann, Mueller Medien: I think it's good they need us at least in the beginning (audience laughter). I came to this conference to get some answers to that. First, we need to feed our sales people. Then there's the question of what the product looks like. I see smaller ads online vs. print with a big price difference. Presenting this on a website means we need some ideas to come around to get a product that works for a small business. I don't see this idea right now as something that can be implemented.
Moderator: To succeed in the local space, will the search engines need to deploy their own outside sales force, or should they partner with someone who already has the feet on the street?
Tom Wilde, Terra Lycos: We're owned by a Spanish phone company (Telefonica), so we already grapple with how to leverage the assets that are already in place. In theory, we all have chunks of assets that we can apply to this to get it done. Print YP is a historical model, and there's a ripple effect as change takes place. Inertia is a big factor.
Moderator: What's likelihood that these kinds of partnerships between search providers and YP companies will be possible?
Tom Wilde, Terra Lycos: Feedback is required. Are Google and Overture friend or foe? There's a high anxiety of partnering and then being "ebayed" where this cute little auction company becomes an unstoppable competitor. It's a difficult dance even though we are the same company. It has not been a slam-dunk.
Paul Levine, Yahoo: The prognosis is pretty good for partnerships with telcos, legal or other category specific publishers, etc. Yahoo's partnerships are working extremely well helping us reach advertisers we couldn't afford to get to. It's good for the telcos as they can offer a Yahoo product as part of a suite of products. It gives them a bigger audience to sell.
Gerry Campbell, AOL: We stand on the edge of a promising future, but it's not completely new. It's all about bringing buyers and sellers together. Can any one company go after this? Who are the right companies to partner with? At this stage, there is no question that companies need to work together to share traffic and advertisers. Partnering is critical to success.
Moderator: Would you partner with a search provider to introduce advertisers to local online search?
Michael Oschmann, Mueller Medien: It's worth talking with them because there are so many threats and opportunities. The big opportunity requires that we not enter a partnership simply as a sales force. We know something about technology and we bring a lot more to the table than just sales. We've spent a lot of money developing a very strong brand in Germany. Those investments are paying off. We own the Yellow Pages trademark in Germany, which is unlike the situation in the USA. We can explore partnership opportunities to reach advertisers that we wouldn't otherwise serve.
Tony Mamone, LookSmart: There's more than just a sales team required. There's a demand from advertisers to get included in searches across many properties, not just one specific search site. The advertiser wants to show up in all searches. In order to do that, you need to partner and develop products that you can sell and offer across multiple search sites. It's not a trivial task, but it is doable.
Moderator: What does a true local search offering look like? Is it structured data, is it searching by keyword, or is there some other variation?
Tony Mamone, LookSmart: There's not one right answer. Users like choice. People will try different types of searches for different types of queries. Each site will need to have it's own personality more than we see today. The user should see different results based on different queries that are typed in. Local vs. non-local results should be separated. Users will ultimately tell us what they want by their actions.
Tom Wilde, Terra Lycos: When you're building a search engine, you're missing most of the structure that exists in an IYP. Getting local information about a web page you stumble across is very difficult to do. Google identifies local tags. Lot of noise in the data.
Maybe there's a way to charge local advertisers to append their web pages with local identifiers. Trying to figure out the geography of the 3 billion web pages out there.
Paul Levine, Yahoo: There are two extremes that are coming together in the middle. Those are web search and IYP. We're incrementally introducing strengths of IYP into web search and introducing some strengths of web search into IYP. Some possibility of geo-tagging web pages. We web search becoming more local. The number one thing that users ask for from us in the IYP is "Tell me if there is an official web site for this company." Making IYP more extensive to find if there is an official website. We've added "search the web" follow-throughs so that users in the IYP can search the web to find if there is more information available out there.
It will be evolutionary rather than revolutionary. You won't wake up one morning and see that local is here and it's absolutely better than anything I've ever seen before.
Gerry Campbell, AOL: What AOL is focused on, and what we think the industry should be focused on is building a "data asset" so that users can have the right experience. There are certain things about a plumber search that are specifically different than a manicurist search. What's the same is that there is a common data set that is searched.
Moderator: Ultimately, is local search a path of technological development or a marketing strategy?
Tom Wilde, Terra Lycos: It's a lot of both. How you connect the user to the merchant is mostly a marketing exercise. How you create local information from the existing 3 million web sites is a technological problem.
Michael Oschmann, Mueller Medien: Both sides need good algorithms and on the other hand, you need to position your product so its perceived to be local.
Tony Mamone, LookSmart: Technology is the issue today. It's only as good as the data you can feed it. Marketing a local search product is a leaky bucket if the experience isn't good. Get the tech down and then work on the marketing.
Paul Levine, Yahoo: You need a good product and then let consumers aware that you've got a good product.
Gerry Campbell, AOL: you've got to get a product that works as you let people know about it.
Moderator: How soon will you see next generation? Overture has said that they will introduce their local search product in the next few months.
Gerry Campbell, AOL: It's happening now. AOL, Yahoo, Google all have a local search component.
Moderator: Some people have claimed that paid search is vehicle for spam.
Tom Wilde, Terra Lycos: Today, everything in paid search is reviewed by a editor and also reviewed by technology algorithms. I don't think there's a giant spam problem. Anytime you launch a new service, you have to take abuse into account.
Tony Mamone, LookSmart: I think that it's ironic that paid has much less abuse than unpaid search for spam. But a short term issue in pricing such that you can go more generic with your search terms and still have a good ROI. Everybody is tracking clicks to see of users are backing out of a bad result.
Moderator: How would a telco with multiple search partners and different products be able to create a scale offering? Wouldn't that put pressure on margins?
Tom Wilde, Terra Lycos: There's a technology challenge. Integrating into search in a deeper level can be difficult. Most telcos don't have the personnel to do the work for deep integration such as xml feeds.
Paul Levine, Yahoo: I actually think there's plenty of margin for the telcos to participate as partners. Less emphasis on price sensitivity as you get deeper into the smaller businesses in the market. There's quite a bit of margin out there and selling multiple products works quite well from a profit perspective.
Moderator: Who owns the customer?
Tony Mamone, LookSmart: Ultimately somebody owns the relationship with the merchant. Billing and collection of data. For now that's the sales force because it's vastly more of a relationship than what we have now. Short term it's the YP publisher. Some of those advertisers will jump to direct relationship as they become more technically savvy. That's not going to happen for a number of years in a big way, and there's so much untapped business that it's not an issue in the near term.
Tom Wilde, Terra Lycos: The near term opportunity is the national advertisers that have local presences. We already work directly with them to present local content.
Tony Mamone, LookSmart: All of us are cognizant that it's early days. What you see today web search is much better than it was but it's far from ideal. We try to determine a set of the best ten possible results for any given search. Interestingly, very few queries today are resolved with the ideal set of ten. Long way to go. Users feel they're finding relevant results and are getting used to a bad experience.
Dick Larkin is Vice President - Internet of TransWestern Publishing, a major independent Yellow Pages publisher with 330 directories in 25 states. He produces "The Yellow Pages Commando News," a weekly newsletter covering local marketing and sales strategies.