Category Relevant Search Results, Query Suggestions, and Advertisements Based on Location?

A couple of days ago, Mike Blumenthal of Understanding Google Places & Local Search asked a pretty timely question with the post Google Local: Are Mobile Signals Actively used in Ranking Local Results? Mike mentioned a post I wrote about Google research on using driving directions as a local search ranking signal.

Mike can add another example of how location may play a role in the rankings of local results.

Street cars in San Francisco.

It’s nice knowing the names of things. For instance, a “Query Stub” is the root (the letters someone types in) of an auto complete search suggestion. At least, we learn that from a Google patent granted a week ago. But that’s not all we learn. Results shown in suggestions responding to query stubs might depend upon your location, at least in searches for business entities.

The patent tells us that someone searching from San Francisco is probably more likely going to search for restaurants or public transportation than someone searching from Yosemite National Park, which has considerably less restaurants or public transportation (No street cars like those shown above).

The category types “restaurants” or “public transporation” are more important in San Francisco that at a national park. Search queries are more likely to be relevant in San Franciso than in Yosemite when they involve one of those categories because there are considerably more entities that might fit into those categories in San Francisco.

There are two different locations that might be important to a searcher. The first of those might be at the actual location of a searcher, as determined by GPS, wifi connections, cell phone triangulation, and microelectronic devices built into something like a smart phone.

The other location might be one indicated by a searcher, such as a location that the searcher has expressed an interest in by specifying it in their Google location settings. (The searcher might be located in San Francisco, but might be searching for results in New York City, for example.)

Imagine that Google has precalculated relevance scores for categories at different locations based upon categories of entities at those locations. Imagine that autocompleted query suggestions for business entity types might tend to favor results that fit best with categories for the location where the searcher is at, or where the searcher express a preference for.

Or advertisements that might be displayed?

Or search results might be reranked based upon how relevant a query might be to how relevant a category is to a location.

Category relevant search results may vary based upon location.

I’ve written about how categories might be used in search patents by Google before, but none of those posts involved location to such a large degree.

I can see how a location relevance category approach could influence each of those.

So, someone searching for “ford cars” in a location where there are lots of Ford Car Dealers might be considered really relevant for a “Ford Car Dealers” category. In a location where there might be a lot of Ford Car repair shops, but little or no Ford Car Dealers, the category that would be the best match for “ford cars” might be repair shops, and not Ford Car Dealers. The patent tells us that this might influence web results, but I’m thinking that it could also influence local search results as well.

The patent is:

Determining relevance scores for locations
Invented by Matt Lewis
Assigned to Google
US Patent 8,407,211
Granted March 26, 2013
Filed: December 16, 2010

Abstract

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for determining relevance scores for locations. In one aspect, a method includes storing a respective plurality of category-location relevance scores for each location of a plurality of geographic locations. A category-location relevance score is based on a plurality of category-entity-location relevance scores for a plurality of entities associated with the category at the location.

A first category-location relevance score is determined for a first geographic location that is not one of the plurality of geographic locations. Determining the first category-location relevance score includes calculating the first category-location relevance score based on a second category-location relevance score for a second geographic location in the plurality of geographic locations and a physical distance between the first geographic location and the second geographic location.

Share

21 thoughts on “Category Relevant Search Results, Query Suggestions, and Advertisements Based on Location?”

  1. I think the local search results must be influenced by mobile search. Most of the people make these searches while on-the-go and that happens with the mobile device.

  2. Hi Ravindra,

    I’ve been presuming that some kind of location based statistical model has been in place for a while, and this is definitely a step in that direction, that works well with actual locations of mobile based searchers, identifying their location based upon things like GPS or wifi access points or cell phone triangulation.

    What I found interesting was the fairly broad scope, from search suggestions to advertisements to reranking of search results based upon category. Funny that the patent doesn’t specifically mention local search in addition to those, but I think that’s been around already in things like how Google calculates location sensitivity for use with how wide a range Google might use in Google Maps for different queries, and the categories that might be assigned to different queries involving business entities.

  3. It sounds like a nice step from Google. making location even more relevant for the user queries.
    We just need to see how much it will affect results as there is some opening here for distortion of what the user truly seeks out when he searches, this may cause the need to be a bit more specific on search queries now on.

    Let’s just hope we will not see ‘business category in a location spam’ now 😉

  4. Hi there,

    I think there is two effect for that type of auto relevant generated query by the google. Firstly, Google makes more specific in SERP. Secondly It makes user less effort as like i search “new car” i get in drop down new car launches , new car 2013, blah blah..

  5. Hey Bill, saw you speak at Local U in Baltimore, good stuff!

    I was reading this post and based on the example you used for car dealers, would it be beneficial to adjust your google plus local categories based on the city? So for example if you own a car dealer in a city without a lot of them should you maybe assign one of your google plus local categories to be like, say, “glass repair” or something? (so long as your business actually does that)

  6. Hi Bill, it seems there are two different intentions that could be seen with local search, “on-the-go” VS “research phase” intentions. For example, maybe something like the archetypal local search “pizza” VS “hotels in [city]”, the second may be either on-the-go or research phase. I suppose device would provide some clue, although you could certainly be doing research on a smartphone planning a vacation. It seems device location would be much more important for an on-the-go search. Thoughts on how Google might differentiate between those seemingly unrelated intentions?

  7. Hi Or,

    I’m not concerned that this will both put more of a burden on searchers and make it easier for people attempting to manipulate search results.

    It presents a more reasonable filter on autocomplete for search suggestions to make it a little more likely that the dropdown suggestions you see are more relevant for the location you’re at. For instance, since I’m in semi-rural Virginia, when I start typing “subway” into my toolbar search box, it doesn’t make sense for the search engine to offer suggestions involving a subway system (there is none), but it does make sense for them to show suggestions related to the “subway” franchise since there are subway restaurants in the area.

    It may boost location relevant results higher in search result than less relevant results, but that doesn’t mean that it will eliminate those.

    It could also make it more likely that advertisements that are relevant for a place are more likely to show up.

    To a degree, this is a geographical statistical model based upon location that makes sense to apply.

  8. Hi Josh,

    Thank you. The Local University in Baltimore was a lot of fun.

    I’d be careful about the categories that I might use on my Google local page to categories that people are actually searching for and stand a good chance of using my business if they see it, and categories that bring in a good return on investment for the amount of work (Do I make more money with “windshield repair” or “transmission repair”? Which do I prefer to do? What earns me more money over the same period of time? How likely is someone going to contact me for one or the other?)

    Chances are that if there aren’t a lot of car dealers in the area, I might suffer some because that category might not be as relevant for the categories I might choose, but I might benefit from lack of competition.

    If I choose “glass repair,” which is as likely something that would fit into a handyman type business category as it would a car repair category, and there are lots of handymen and very few auto shops, Google might associate “glass repair” more with a handyman type category and less with a car repair category. I would also suspect that “windshield repair” is more likely what people would search for, and it still might be the most relevant category for a car repair place even if there aren’t many locally.

  9. Hi Matt,

    As I noted in the post, Google could consider two different types of locations – the place you’re at presently or the place you may journey to. For those places you’re planning on traveling, you could include geographic terms in your query (a likely approach for many people doing that kind of search), or you could even change your default location settings in Google. That’s something I would consider doing if I was planning on spending a week somewhere, and I wanted to locate things like restaurants to eat at.

    The patent does mention searches where you indicate the location in a manner like one of those, or one based on the location that you’re at presently.

  10. Hi Rashmi,

    The patent also notes that it might show different results in sponsored ads that it shows for a search results page based upon the location of a searcher. If I search for “subway” 50 miles west of Washington DC, I’m less likely to see advertisements for the metroline subway around Washington DC, and more likely to see ads for Subway sandwiches.

    So, it could impact dropdown search suggestions, search results, and sponsored advertisements.

  11. Hey Bill,

    You got a very well-written post. I guess Google has done it again. If this is to be happen, then it will be much easier to travel or at least know the place you’re traveling at. For example, tourist from other countries pays a visit to some of the most beautiful cities in the world. With this kind of application which can be access through their android power devices, tourist will have more time to search for many tourist destination places. Thanks for sharing this post.

  12. Over the past month or so, I’ve noticed that the location of the searcher certainly influences the results he/she sees from Google. Many prospects register on our site with keywords that don’t include a location (such as “condos for sale” or “townhouse communities.”) Based on their IP address (if they’re in the Charleston area), our website shows up in their results. This is a simpler example of what you’re describing, but I think it still shows that location is getting factored more and more into Google results.

  13. Bill, great article. I have noticed that we are receiving more traffic from Google through our Google Places. It is nice to see that Google is putting more emphasis on Local results when a generic keyword is entered. Your post reminds me that I need to update our local listing with more up to date pictures as well as responding to reviews. Thanks for the great resources that you have posted.

  14. I have noticed something strange about Google lately. I am an SEO consultant/Marketer for a company in Ecuador, based out of Canada. When I recently put on a map from Google Maps showing the location of the company in South America, I was stunned when I seen the map a week later showing the address of the company in Canada instead of the address in South America. Was this a hiccup from the Gooogle spyder, or something else. This was not a problem before Google made their update to their search engine.

    Has anyone else noticed anything strange while doing SEO?

  15. Bill

    One of the questions that always pops up in Google local search is: Why does Google show a 7 pack for CityA + service but not for CityB + service?

    Do you think that this relevancy patent might play a role in explaining that outcome? IE the local pack is not considered relevant for a particular geography.

  16. Hi Mike,

    It’s definitely one possible reason why Google might display a seven pack for a particular service in one city but not another.

    When I search for [tour guide washington, DC] I do get a seven pack, and when I search for [tour guide warrenton va] I don’t.

    Of course, it’s also possible for Google to do things like click experiments to see if people click upon special results when those are shown in results, also.

  17. I don’t completely agree with one of the comments. They said that most of the local searches might be preformed by mobile devices. I worked for a long time in an industry where people searched based on destinations. Most of those searches were done by desktops from the analytics that I did back then. I am guessing that Google aggregates everything and bases the patent on both types of searches.

  18. What I find puzzling is how well Google might be able to untangle the relevance of wanting a local result. Let’s say I want to buy some shoes from a “shoe shop”. If Google brings me up lots of local stores I’m gonna be happy because I probably want to go along and try them on. Local results make sense for this query. However let’s say I want to buy a “large wrench”. In this instance I’m not bothered about local results, I don’t need to try it on, I know what I want to buy so I would much prefer national results because I just want the best price for a commodity of this type. People don’t always want local for commodity purchases, in which case a set of local results is not a useful things to return to the user. Even if I live in the wrench selling capital of the world, doesnt mean that the local results are the most relevant. All the local firms might be terrible, it’s just there are alot of them…..I need to lie down now…..

  19. Hi Bill,

    Great Post! Well what can I say, Google has done it again!
    This will defiantly make it much easier for tourists in a local area to find travel destinations and tourist attractions, restaurants and bars in the area for which they are visiting.

    Fantastic post & thanks for sharing! Keep up the GREAT WORK!

Comments are closed.