I reported in last Wednesday's All Points Blog that Yahoo! had ended its relationship with deCarta, its underlying mapping technology provider, to switch to an in-house developed solution. Directions Magazine received a call from deCarta on Thursday to elaborate on the change. We spoke with Yahoo! on Tuesday about the decision.
deCarta CEO Kim Fennell reiterated several times that Yahoo! came to a business decision that was not based on any sort of dissatisfaction with deCarta or its products. What drove the choice? Fennell explained that the driving force behind Yahoo!'s decision was the growing importance of local search to its business. He shared that Yahoo! had made the decision to move to an in-house platform a year ago. Yahoo!, he said, wanted to "control its own destiny" and not be dependent on a third party in this important market segment.
Local search, Fennell noted, is a $30 billion market and everyone wants in. (Much of the revenue is in advertising sales.) Google, Yahoo!, Ask, and Microsoft are all in. They are disrupting the "Internet Yellow Pages" (IYP) companies, who, in turn, are becoming potential clients for deCarta, he said.
Fennell described Yahoo!'s move from deCarta to "in-house" in 2007 as parallel to the decision five years ago to move from the MapQuest service solution to deCarta's platform. (At that time deCarta was known as Telcontar.) Yahoo! wanted a platform that could be hosted in-house and be customized with the Yahoo! look, feel and brand. The company wanted to "raise the bar" and build its own functionality on the deCarta platform. Now, Yahoo! is taking the next step.
If dependency was a concern and there was no issue about functionality, why didn't Yahoo! acquire deCarta? Fennell could not confirm or deny whether such discussions occurred, but he could confirm that a sale was not in the current plan for the company. And, he made it clear that while the company would have preferred to keep Yahoo! as a client, the contract was worth less than 3% of revenues.
Another way to look at the decision is to note that Ask and Google (among others) use the deCarta platform. Does the Yahoo! decision to take its solution in-house help distinguish it in going after the local search market? deCarta staffers all agreed that it gives Yahoo more control of their destiny. But, Fennell commented, that puts deCarta in a good position to make its offering available to specialty sites like Zillow.com or Hotels.com that need to keep up with the rapid advances in mapping but ultimately have expertise in real estate and hotels, respectively, not in mapping. Such companies would not have the resources or interest to take their mapping solutions in-house. Who would have the resources and interest? Fennell pointed out that certainly Google, Microsoft (which already did) and Yahoo! could, but few other companies really could or would.
Jeremy Kreitler, Director of Products for Yahoo! Maps offered a different reason for the change. He noted that while local search is important to Yahoo!, it was not the driving factor in the decision. He explained that Yahoo! wanted more control, the ability to support global expansion, and that it was difficult to develop some features and functions on the platform. Yahoo!, he said, wanted to develop an "offering based on what users had asked for a unique, valuable and best in class solution. [The company] wants to distance itself from the competition. When everyone is on the same platform it is difficult to do that."
Kreitler explained that making the move took "a while, even with our talented people." He also highlighted that the current offering is just a "starting point" and further enhancements could be expected. He said that in internal comparisons the new version faired very well in comparison to the previous versions and competitors. In particular, Kreitler noted an increase in "accuracy." That, he clarified, referred not to data, as there are limited new data in the release and NAVTEQ continues to provide street data, but rather to "enhanced accuracy in geocoding, more efficiency in routing, and more readable maps." In short, the algorithms and presentation are "smarter."
I asked about the technology used, specifically about the use of open source tools in the in-house implementation. Kreitler reviewed Yahoo!'s history in open source as an innovator with Apache and MySQL and explained that his team had looked at open source code for the mapping application. "We didn't see a good fit. We found that most solutions were geared toward business intelligence for small and medium businesses." So, what is under the hood? Yahoo! says its a blend of technology developed internally and to a lesser extent, licensed code it acquired. As standard practice, Yahoo! does not disclose the nature of all its business relationships, but given the scope and complexity of this project, said it is common practice for them to work with several partners; most of which now fall on the data side of the equation.
How will the new platform tie in to local search? In short, he explained that better mapping simply means better local search. According to Kreitler, it also means better maps for other Yahoo! properties. "There are already 15 properties that tie into Yahoo! Maps and others likely will in the future." Among those integrating Maps are Yahoo!'s travel, real estate, and mobile sites, as well as the geotagging tools in Flickr, Yahoo!'s imagery sharing site. When pressed to highlight what's coming, Kreitler spoke to the importance of the Yahoo! Maps Application Programming Interface (API). "It's very important since we can't think of all the possibilities nor do we have the resources to build all the apps." I noted how well Yahoo! Pipes integrates with Maps and Kreitler explained that the Maps team will be highlighting Pipes more and more in the future.