March 19, 2007
Last month there were reports of errors in Google Maps listing for Duke Medical Center phone numbers. Not only were the numbers often times wrong but Google Maps was listing as many as 5 phone numbers for a single facility, few of which went to the central switchboard. To Google’s credit the Duke Medical Center situation was resolved rapidly on an individual basis.
When an additional report surfaced in Google Maps for Business Owners from NYU Medical Center, I decided to investigate the depth of the issue and see if was isolated to large teaching centers or was more widespread throughout the medical listings.
The report from NYU reported not just multiple and wrong phone numbers but wrong map locations as well. Google’s new feature in the Local Business Center, allowing the movement of Map markers, will facilitate some corrections about location but not the problems with phone numbers. Most medical centers have difficulty complying with Google’s protocol for record correction as they have multiple mail stops making delivery of the PIN card unlikely and multiple phone lines making a call for verification almost impossible.
The medical center phone data is in critical condition and in need of intensive care. .
Only 2 of the 13 markets did not have listings showing 4 or 5 phone numbers. 18% of all listings showed 3 or more phone numbers despite low verification rates through the Local Business Center. And if you live in Fargo, ND it appears by Google’s account that there are more medical phone numbers than there are doctors.
When I found 5 phone numbers listed for a hospital, I spot checked by calling some of the numbers to see where they went and if the answerer had received erroneous calls. In very limited tests, the numbers went to a department rather than the main desk facility listed and noted receiving a fairly large number of wrong number calls. I think it is safe to assume that any facility with 4 or 5 (and likely 3) numbers listed probably has numbers not for the main facility or that are in error.
The fact that none of the medical facilities with 3, 4 or 5 phone numbers listed had yet to claim the record in the Local Business Center shows: 1)that there is a possibility that they will get cleaned up and 2)that it is a ways off. Another (perhaps more real world) test of the quality of this data would be to see if the single number that Google picks for these records in the Local OneBox is the correct one.
March 16, 2007
Here are some recent announcements that have not been covered widely that I found of interest:
NavTeq goes 3-D
Summize: Enhancing Search With Heatmaps
Maps as art and politics (nothing to do with local search per se but interesting none the less)
March 15, 2007
Ahmed Farooq of iBegin.com contacted me several weeks ago to alert me to a new product that his company is releasing today: iBegin Source, a free and low cost source for business data in the 50 states.
Traditionally business data from InfoUSA and the like is expensive and very restricted in its use. For example, InfoUSA quoted $101,038.34 for every business in NY State as opposed to the $1000 for the iBegin commercial data set for New York. While InfoUSA includes more and different fields in the data, for many uses the geocoding available with iBegin may be more valuable.
iBegin Source is making this local data available for each of the 50 states. From iBegin Source’s website:
- FREE download for non-commercial usage
- Commercial license is only $1000 for a state or $40,000 for the entire USA. Other data brokers can cost more than $500,000
- Automated purchase. No sales team to go through
- Data is updated constantly. Includes daily, weekly, and monthly data updates
- We have 10,820,453 total business listings. Already cleaned and de-duped
- Commercial license includes geocoded addresses
The availability of free or cheap local data that is updated regularly has the potential to shake up not just the internet yellow page business and local search but direct marketing as well. To quote Ahmed: “We want to help promote enthusiast and hobbyist sites (just look at what happened with mapping applications when Google released the Maps API).”
Here is an interview that I did with Ahmed over the past few weeks that provides more insight into his service:
Q:Tell us about your company and how you got into local search
The parent company is Enthropia Inc., a webdev firm based in Toronto. We are self-funded, over four years old, and we build our own sites (no client development).
We got into local search because the current crop wasn’t good enough. From massive errors in data to slow searches, it was a headache to find anything near me. Canadian local search is especially horrible. We didn’t want to take a shotgun approach, covering all of Canada/US. We opted for a city-by-city approach (ala CitySearch).
Describe your new service to provide local data
Local business data is expensive. The data itself is full of duplicates and errors. I remember processing 34,000 records for a city and ending up with only 8000 unique records. Brad Fled had an interesting post on how bad local data is, and how the suppliers provide of no way for direct updates.
So iBegin Source does four things differently:
1. Perpetual license. Once you purchase our data, you can use it for however long you want.
2. Cheaper data. An entire state is only $1000. The major data brokers (that everyone uses) are roughly 300-400% more expensive than us. Some are high as even 1000% more! We want to help promote enthusiast and hobbyist sites (just look at what happened with mapping applications when Google released the Maps API)
3. Open system for updating. Anyone can submit an update, and we also have a trackback system for automated updates (akin to what Brad Feld was suggesting). All of it revertable just like Wikipedia. No more closed systems.
4. Geocoding comes included. Six decimal accuracy and major intersection included. Simplifies the entire process.
What do you think will be the impact of making this data available?
The entire idea is that helps launch new local-oriented sites. If I wanted to setup a local site right now, the cost of the data is a major barrier. With iBegin Source I can self-fund my project.
We also intend on becoming the centralized place for local data. On the iBegin city sites we have received thousands and thousands of updates, of which a fraction of one percent were incorrect. Why not just give users the power to do the changes themselves? Worst case situation: we do some reverts.
As a partner in a small firm that relies heavily on technologies that support and enhance telecommuting to maintain customer contact, I found three (totally unrelated) announcements of interest:
Microsoft acquires TellMe. I have been a big fan of 877-520-Find as a safe and productive tool for cell phone calling. It demonstrates effectively how voice (even though frustrating 10% of time) can move local search data easily into the mobile world. It avoids the DWT (driving while typing) offense that seems likely to kill people and is less frustrating than talking to directory assistance in a foreign country. The TellMe acquisition will put this type of service on the front burner as Google, Microsoft and others compete for the next generation of mobile search.
Gizmo SMS has rolled out a new service that allows you to send SMS text messages from your browser to mobile phones around the world, for free. I have always thought that texting was an effective business communication tool but just never could bring myself to pay for a web to phone service. It is simple, effective and even works from dial up via an old browser.
David Pogue of the NY Times writes (reg. req’d) of a new service that provides â€œOne Number That Will Ring All Your Phonesâ€. A new service by GrandCentral.com (free for 2 numbers, $15/month for 6 phones) will ring all of your phones simultaneously and keep all of your messages in a single web based voice mail box.
All of these services integrate existing mobile phone technology with the internet in a way that increases the value of both without requiring new hardware for either. They all could be used for delivery of ads and all offer a glimpse of the immediate future of integrating data with our cell phones.
March 10, 2007
locally type(d) has a write up of the new Yahoo Local/Maps suggestion board where users can suggest, comment and vote on the best ideas for improvement.
If you have a idea for Yahoo Local you canmake your suggestion here.
I find the differing approaches of Google and Yahoo to map corrections/suggestions of interest. Yahoo one the one hand allows the more liberal, free for all approach allowing general users to add information to the maps but the more strict and less volatile approach to suggestions of voting. Google on the other hand, has a strict approach to map information input, allowing only the record holders to make changes in the Local Business Center and the much more freewheeling approach in their Google Groups for Maps for Business owners area.
While I will miss that Google Groups give and take in the Yahoo approach, I am sure that it leads to better employee relations. I am convinced that each morning, Map Guide’s Brian and Jen go to work and have this conversation: “Jen, I think it is your turn to respond today.” and Jen says: “No, no, I insist, it’s your turn.”
We’ll have to see if it leads to less material for my blog.:)
March 9, 2007
Thursday, Google introduced the ability to add custom attributes to a business listing at the Local Business Center. It is of interest that the default custom attributes change by industry. The choices for additional detail fields change as well. For example with a service industry the first field is Price, in the restaurant business it is Ambiance and for a Physician it is License Information.
You can see the default choices in the samples below. Google is essentially creating a public means to access the “schema” for each industry and to also help define it. Are we seeing the beginning of the “semantic web” race?
Greg Sterling’s recent post: Print Yellow Pages: Whatâ€™s the Real Story? is one of the best summaries of the issues and realities of the print Yellow Pages and their positioning vis-a-vis the internet yellow pages and local search.
I have fallen into the “detractor” camp for a number of years but have recognized that my personal point of view does not a trend make. Greg’s conclusion:
“Itâ€™s simply untrue to assert that all is well in print yellow pages land. But itâ€™s also incorrect to call the medium dead……paradoxically, as consumer usage continues to migrate online the bulk of the revenues will remain offline for some time.”
Speaking of schema and the semantic web, the NY Times today had an article on Freebase(registration req’d), from start-up Metaweb that has the goal of “trying to create the worldâ€™s database, with all of the worldâ€™s information,â€ based on the ideas of the semantic Web.
According to the NY Times, “since it could offer an understanding of relationships like geographic location and occupational specialties, Freebase might be able to field a query about a child-friendly dentist within 10 miles of oneâ€™s home and yield a single result.”….
â€œItâ€™s like a system for building the synapses for the global brain,â€ said Tim Oâ€™Reilly, chief executive of Oâ€™Reilly Media, a technology publishing firm based in Sebastopol, Calif.
Despite the articles fawning, messianic tone, it is interesting to me that the these ideas are now receiving coverage in the mainstream press.
Google slayers and purveyors of “all of the worldâ€™s information” will come and go. Some will survive and offer interesting developments and one might even one day unseat Google (and it might just be Freebase). Regardless, this road will be long, winding and interesting both in the technologies and the competitive battles.
Business Week wrote a piece recently (03/08/07): Where Search Stumbles criticizing most of the local search engines. Greg Sterling has repeatedly pointed out that accuracy is the potential achilles heal of Local and there have been plenty of complaints elsewhere and here about quality and about complexity. But as I tell my relatives: I love you anyways.
I by no means find the quality acceptable, although I have gone on record with the opinion that it is satisfactory and will improve. I recently, though, read a post called “Its the Schema, stupid!” by a search researcher for IBM that helped me understand the bigger issues. I am not a database expert and have limited understanding of search schema or the semantic web. However his point that what is exciting about local search is that the searcher is looking for something real, struck a chord. The searcher is trying to find a pizza parlor or a car dealer as opposed to a document in cyberspace that general search might return. And that only by using a structure (schema) that is based on the real world would you ever be able to find it.
This brought home for me both the elegance and complexity of what Google, Yahoo & Ask are attempting to do with local search.
In Google, an organic search returns ranked & relevant results from a selection of web pages. They can not be judged as right or wrong just more or less relevant to the searcher and only in the aggregate.
In Google Maps local search, on the other hand, Google is attempting to return ranked, relevant AND right results. The relevance can be measured from the local searchers point of view, the rank can be disputed by the local resident and the rightness (or accuracy) can be verified to see if Google properly captured the correct phone, street address, category, neighborhood, quality, recognition of a hopefully open local business. It is judged in the very specific.
March 8, 2007
Another report of large medical center problems with Google Maps:
== 1 of 1 ==
Date: Thurs, Mar 8 2007 9:22Â pm
From: “Michael” Â
I have the same exact problem. I am the web director for NYU Medical Center.
I have the correct address and phone number listed in over 20,000 pages on the footer.
Google maps does not have an elegant mechanism for validating changes. With large institutions, mail stops can be very difficult so the post card method does not work. Also, our call center is analog, so they can not validate there. How about a validation tag on our web site?
that would seem to be the most logical.
This has become a serious problem for us, as we have patients
literally showing up in the wrong locations when they are scheduled for surgery.
I have sent several emails to the maps group, to no avail.
Bart, if you find a solution i would love to hear it. I would love to know where the mapbot is getting its data from. An XML document on the root of my server listing the correct addresses and numbers would make the most sense to me.