GPS to Correct Google Maps and Driving Directions as a Local Search Ranking Factor?

I love local search. It follows many practices similar to Web search, though different often in ways that do reflect an attempt to map the real world. Google’s Streetview cars are a little like Google’s webcrawler Googlebot. Instead of collecting URLs for Websites, Google Maps collects addresses to associate with businesses, nonprofits, government offices, parks, landmarks, and many other destinations. It has its own challenges as well, such as the streetviews car being turned away at sentry guard booths for military bases, or not being able to drive down “private” roads. Google Maps also can’t use latitude and longitude coordinates in places like China since their export and use is classified by that country as if they were munitions.

A mobile phone screen showing the location of a searcher, near the Googleplex

I’m also often frustrated by local search. Driving directions from Google often begin by telling you to go “east” or “west” on your first turn. I’m not Mason or Dixon, Lewis or Clark, and I don’t carry an in-car compass with me when I drive. I often have no problems with the directions other than that, for the first 99% of the trip, and then have problems with the last few hundred feet.

Updating Google Maps

Sometimes places don’t exist where Google Maps say they do. That can happen when places close and it’s not reported to maps, or places move and don’t update their address information on the Web, or sometimes places just aren’t where Google Maps say they are. I tried to find a local restaurant last year, and drove around in circles until finally giving up. I later found that the place had closed, but wasn’t sure of the right way to try to report my inability to find it to Google when I was driving around, since I wasn’t sure that it had closed at that time.

A Google patent application published last week describes a solution to reporting incorrect addresses that uses GPS and Social Networking to sometimes provide new addresses from others who have tagged the location of a place. I could have really used this about a month ago, to try to find the right location for a place I was trying to visit.

It’s costly to send “expert observers” out to update mapped locations, which is the traditional approach that map makers often employ. Because of that, some map makers don’t update location information very frequently, especially with printed copies of maps. With online mapping services and portable navigation services though, we expect up-to-date and timely information when it comes to the locations of businesses.

Many online mapping services provide for individual users to make suggestions for corrections to maps, but those often have to be reviewed and approved by an expert reviewer. Some of those systems might accept changes from users without review who have had a large number of suggestions approved in the past, and have built up some level of credibility.

But what if a user of a mapping service arrived at a destination on a map, and the place that was supposed to be there wasn’t, and they reported it on their mobile phone while they were there. As in the screenshot below from the patent filing, those systems can be tracked a number of ways, including through the use of GPS or cell tower triangulation or wifi location reporting.

The same map as above, but with a notification that wifi data is being recorded for location

Imagine that not only could you report a missing destination, but that you could send a correction into the map, have local contacts of yours queried for the right destination, and possibly find the place you were looking for while also helping to update the mapping service.

The images below are screenshots from the patent filing, that do a decent job of showing how such a system would work:

The map with a dialog box, asking if the viewer is attempting to correct an existing location

The system might help you try to find the place you are looking for by displaying a list of destinations nearby:

A list of destinations near the location, that the viewer can select from, for the place that they are trying to correct

This system allows you to enter information about the destination that you are at, including title and a description as well:

A screen allowing for a description to be updated

You may see information displayed from other people who have used the mapping service and have tagged the location of your intended destination at a different place, or who are contacted by the system and asked to give their input about the place you’re trying to visit.

A screen showing the location tagged, and a message stating that 5 contacts claim a different location for the place being searched for

You then have the ability to agree or disagree with the other people who have tagged a different location for your destination.

The map interface provides the viewer with a chance to agree or disagree with their contacts claims of the other location.

The patent application is:

Trusted Maps: Updating Map Locations Using Trust-Based Social Graphs
Invented by Chaitanya Gharpure, Charles L. Chen, and Tiruvilwamalai Venkatraman Raman
Assigned to Google
US Patent Application 20110238735
Published September 29, 2011
Filed: March 29, 2010

Abstract

A system and method for updating and correcting the location of geospatial entities, the method comprising receiving at a server from a mobile device operated by a first user, a proposed location for a geospatial entity, the proposed location determined by a wireless location system, and based upon a current location of the mobile device; providing information about the proposed location for the geospatial entity to a first plurality of other users; receiving votes from the first plurality of users as to whether the proposed location is correct and responsive to the received votes, determining whether to update the location information for the geospatial entity.

The patent application describes sending out requests to people whom you might know through a social network. The patent mentions Orkut, though something like this might be tied to the Google Places recommendation service, or possibly even Google Plus at some point. The “trust metric” mentioned in the title to the patent filing might be the number of votes agreeing with a proposed location, or a certain percentage of votes. Other groups of users who may not be contacts of yours may also be asked about the location as well.

Driving Directions as Ranking Signals

I ran across a paper last night that describes the possibility of including requests for driving directions as a ranking signal for some types of businesses. The paper is HyperLocal, Directions Based Ranking of Places (pdf).

Part of the abstract from the paper tells us:

Specifically, the paper proposes a framework that takes a user location and a collection of near-by places as arguments, producing a ranking of the places. The framework enables a range of aspects of directions queries to be exploited for the ranking of places, including the frequency with which places have been referred to in directions queries. Next, the paper proposes an algorithm and accompanying data structures capable of ranking places in response to hyper-local web queries. Finally, an empirical study with very large directions query logs offers insight into the potential of directions queries for the ranking of places and suggests that the proposed algorithm is suitable for use in real web search engines.

Of course, data like driving destinations as a potential ranking signal is likely much more appropriate for listings of businesses that people will visit in person, as opposed to businesses with a local presence that might provide a service or deliver goods in specific areas.

If you’re interested in how local search works, this paper is worth spending some time with.

Share

31 thoughts on “GPS to Correct Google Maps and Driving Directions as a Local Search Ranking Factor?”

  1. Every so often, I suddenly am struck by how Orwellian our world is with the rapid advancement of technology. I think these patents are very interesting and would correct some mistakes that I also have encountered with Google Maps. However, I couldn’t help taking a step back and thinking about how these devices that we keep in our pockets are feeding information to a massive information aggregate about where we are, who we’re with and what we’re doing. And occasionally the info-aggregate can’t figure out the precise point where a place is located so they make our devices morph into cartographers for them and that’s what today’s patents are for. Anyway, apologies for the off-topic and paranoid rant, but thanks for covering this Bill! This site has become something I look forward to reading everyday!

  2. Note to self: Pitch new realty show concept to cable TV executives about the misadventures of the Google Streetview car drivers.

  3. Hi Corey,

    Thank you.

    Both Apple and Google have been the target of criticism over location data collection processes, and have insisted that they aren’t tracking individuals, but rather are using the data in ways that can help improve the services that they offer. I don’t know if you caught my post Crowdsourcing Behind New Apple Local Search Patent, but that describes a way that Apple might use location data to rank sites in their maps applications.

    It’s very tempting to think that either company would collect as much data as they could get away with, to use in ways that does seem pretty orwellian.

    Google does appear to look at GPS readings to improve Google Maps, and to derive real time traffic estimates, and they are mapping the word with us as data points.

    Of course, I would be much more likely to trust someone who is telling me that some place is not at the address my mapping system says it is, if I could tell that they are actually at that location from their GPS coordinates, or through cell phone triangulation or wifi location. But it is somewhat frightening that they might do that.

  4. Hi Joel,

    That might actually make an interesting series. :)

    I’ve read some rumors about a reality TV series that would focus upon people working in social networking sites, and honestly I’d rather watch a fictional series about a Google Streetviews driver and his or her adventures.

  5. Hi Luis,

    The first screen in the post includes Google Latitude features, and there’s a Check-in that Google has developed which is similar to 4Square:

    http://www.google.com/latitude/checkin/#quiznos

    This patent filing really doesn’t focus upon those however, but rather explores ways to try to make it easier to correct addresses in Google Maps. The premise, that people who are actually at a specific location where a place is supposed to be according to a map, but isn’t, and can get information both from people they are connected to, and people who they don’t know, to verify a different location, might just work better than relying upon people to report an incorrect address from their home desktop computers.

    One of the really interesting things about this patent filing to me is that it focuses upon a mobile application though, rather than desktop computers. I think that’s definitely a sign of things to come from Google.

  6. This could be a great opportunity for Twitter to add an address section on their profiles. Facebook already has a section for a company’s address and website. And hopefully Google+ will let companies join in on the fun as well.

  7. Hey Bill, I’m fan. Been one for a while.

    I think crowdsourcing has it’s place, but I think it’s also very dangerous…especially in local search. I think the search engines like Google need to provide both the original info they have/had with community edits.

    It’s examples where Mike Blumenthal shows how with 2-3 community edits a Places Page can marked as closed.

    However, on the flip side, I get it. For places that don’t exist yet on Google Maps or update to date (real time) traffic reports from community members this stuff can be very useful.

    I just hope they don’t use it too much rank businesses. Maybe a minor signal would be good.

  8. Hi Hanif,

    I don’t think that I’ve seen anything specifically from Google, or in Google patents that would indicate that KML file coordinates help with local rankings.

    But I do expect there to be tighter integration between Google Maps and Google Earth in the future, so it’s possible that might be something that does matter when and if that happens.

  9. Hi Matthew,

    Thank you. Crowdsourcing does have issues, and it’s a topic that I’ve been looking into more and more. I think that when it’s implemented wrong, it can be more harmful than helpful.

    What makes the approach in the address correction patent interesting is that they look at the geographic coordinates of the person doing the reporting of an incorrect address to see if those are from the address that the place they are attempting to correct is located. If it is, they appear to give more trust to it, but not complete trust. They still query others, including both people the person doing the reporting might know, and to make things more trustworthy, people that the person doing the reporting might not know.

    But the patent involving corrections has nothing to do with rankings.

    The paper that I wrote about at the end of the post involves how driving directions might be used to influence rankings. To a degree, it makes some sense to look at those types of things by Google, and explore whether or not they might be useful in ranking places.

  10. Hi Ryan,

    I believe that Twitter created a geolocation API back in 2009, and I’ve heard some about”Twitter Places” but they don’t seem to have done too much with it. Here’s one of the latest stories I could find about it:

    http://mashable.com/2010/12/08/twitter-places-data-sets/

    I don’t know that they will add address sections to profiles.

    Google Plus is supposedly working on business profiles, but it’s been somewhat quiet on that front recently, too.

  11. So do you think that google maps usage effects google places rankings?

  12. Hi Mike,

    I think that there’s a good possibility that it will in the future if it isn’t being used right now.

    For example, if Google showed Place results within Web results for certain queries and no one ever clicked upon those, they might not be seen as a good match for those queries. Of if Google showed “payless shoes” in local search results on a query for [*location* restaurant], where the actual location was zip code or city name or some other meaningful location information, and no one clicked upon it after a very large number of views, it might not rank very high. (And yes, I’ve seen payless shoes show up in that search locally, since one of the categories they are associated with is “restaurant supplies.”)

  13. This is actually a very good and useful concept. This provides a huge unpaid work (enthusiastic) work force who will improve the quality of Google Maps in a very short period of time compared to a paid work force. I wonder if any of the existing Satnav companies will reflect the free nature of updates in their pricing.

  14. Hi James,

    It seems like Google has recently started getting more users of Google Maps involved in checking upon and updating Places information. See:

    http://blumenthals.com/blog/2011/10/26/google-places-community-edits-of-claimed-listings-are-back/

    There is a very real risk that some people participating will abuse a system like this, but I still think that it’s worth exploring on Google’s part.

    I’m not sure that the satnav companies really have much to do with this at all.

  15. Hey Bill.

    Great post! I really appreciate your research and opinions.
    I firmly believe that dictating a business ranking by the amount of times that visitors seek driving directions for that business has its benefits. If this does go into effect, then many of my clients will be rewarded greatly!
    I only hope that Google has implemented sufficient measure to ensure that people (SEO and Internet Marketers) cannot “cheat” the system.
    Keep up the great work ;-)

    ~Scott

  16. Hi Scott,

    Thank you. It makes sense for Google to experiment with this type of information to see if it really does help them when ranking local search results.

    I suspect that they would look closely both at the data, and at the ways that people might attempt to manipulate search results when deciding whether or not they would use this data. It’s possible that they might look at other signals as well, in conjunction with driving directions to see if this data is useful. For instance, do people also print out those directions, do they send a link to others, etc.

  17. Hi Bill.

    This is very interesting. It reminds me a bit of MapMaker (http://www.google.com/mapmaker) – but on mobile. BTW, MapMaker Pulse can be interesting to watch (http://www.google.com/mapmaker/pulse), as you can see actual edits being made and published.

    However, I am a bit concerned about crowdsourcing as it will be abused at some stage. This already happens: I can give you actual examples of a large accommodation booking company hijacking small hotels/motels listings on Google Places here in Australia to replace website URLs with their own…

    Trouble being – with so many users relying on those indications, how will Google react when the abuse level has become intolerable?

    Cheers

    =Bruce=

  18. Hi Bruce,

    From the description of a presentation by a Google Software Engineer a couple of weeks ago, it sounds like Google may be using the CrowdSensus algorithm (which I just found out about recently) with Google Mapmaker as well:

    Crowdsensus: Building reliable Google Maps from data provided by unreliable sources

    I’m not sure how it might work with the edit corrections process tied to this patent, but I think there might be a connection. It’s worth looking into to see how Google might attempt to try to handle some of the problems with crowdsourcing.

    I hadn’t heard about pulse, but it’s kind of fun to watch. Thanks for pointing it out.

  19. I think the problem you had of looking for a restaurant listed on google local that isn’t there any more is very common. Obviously when a local business opens one of it’s priorities is to get listed everywhere so it can be found. However when they go out of business the last thing on their minds is removing all those listings. Because of this I thinks it imperative that everyone has the ability to report incorrect locations otherwise over time the the amount of incorrect location info will become huge.

  20. Hi Matt,

    I agree. Most businesses don’t have a lot of incentive to contact Google or web-based business directories that they are closing, and they rarely do.

    My incentive to report businesses that are closed is that I’m hoping that others will too, so I don’t set off to find some place I’ve located in Google Maps, only to arrive at a building with boarded up windows, or find a new business has replaced the old one. If there’s incorrect or out of date information about businesses around where I live, that can negatively impact the local economy around me.

  21. Sounds like this algo could be improved: As long as you are still outside it is just approximate.
    If someone uploads a video / picture with the appropriate tag from inside the location, they could use the gps data.

    But these solutions have some flaws too. How do you handle the location of a state park? Quite some area – and all locations would be right. So you need to get down to more granularity than that. There is some work going on on in-building location services, but setting markers at East or North entrance? Would be curious how they treat ‘fuzzy’ locations, perhaps with a calculated center of all location data above a certain reliability index?

  22. Hi andreaswpv,

    We can’t be sure that Google is following what is described in that patent 100%, but chances are that when they come up with processes like these that they add things and modify them when they start coming across the kinds of real world problems that you’re describing.

    Sending pictures is an interesting idea. I haven’t checked to see if Google still has a local used car dealer listed, but the dealership was bulldozed and a Wawa is in the early stages of being built. A picture with the right geo meta data associated with it would be one way of confirming that the place was closed. :)

  23. Hmm whether or not google places is any good is all theoretical for me as I have been trying for months and months to get a duplicate listing removed, which effectively makes me invisible.
    I currently get zero traffic as unless you’re zoomed in to the office, it doesn’t show up due to duplicate content. If you’re ever going to put a business on google maps, whatever you do make sure you get it right first time as editing it is what caused my problem.

  24. @RDR:

    Go to the “Places” listing and claim it. Then log in and go to the listing manager that shows your listing(s) — there should be two of them. Simply delete the duplicate.
    Some times there will be a listing that has yet to be claimed. I ran in to this as well. Find it, claim it, delete it.

  25. Hi RDR and Scott,

    Thanks for the advice to RDR, Scott.

    I’ve been in that situation where there were a couple of duplicate listings, and the thought of deleting one of them is a little intimidating. :) There’s just no certainty what impact that might have, and not a lot of guidance from Google in terms of what to do in a situation like that.

Comments are closed.