Google was granted a patent this week to enable them to check whether people are following recommendations in search results to visit geographical locations – to see how many people actually visited those places. And it’s probably good for many businesses to have Google recommending that people visit them.
For example, people living near a certain restaurant may be recommended as a place they might like to go; and the patent determines whether or not people may be following those suggestions.
The patent provides a summary of the steps that it takes in ranking recommendations based upon whether people actually visit the places being suggested:
In some implementations a computer-implemented method may be provided that includes the steps of: providing a suggestion to one or more users, where the suggestion may be indicative of a potential user activity at a given geographic location; determining locational data for the one or more users; determining, based on the locational data, how many of the one or more users visited the given geographic location after receiving the suggestion; and adjusting a ranking of the suggestion based on a determination of how many of the one or more users visited the geographic location after receiving the suggestion.
This approach may look at particular activities that determine whether or not such a system might use. They can include things such as:
Check-ins – Determining if many people who receive such suggestions are actually visiting the geographic locations recommended
Cellular Tower Signals and Wi-Fi Signals are available in an area that can be used to check upon visits to those geographic locations.
For such ranking suggestions to be demoted, the patent says that a certain threshold number of people who received recommendations for a specific location must have failed to visit that location.
The visit may be required to meet a visit duration time to the geographic location, and the impact on the ranking may be based further upon the visit duration time.
The patent in question is this one:
Inventors: Andrew Tomkins and Shanmugasundaram Ravikumar
United States Patent 9,563,641
Granted: February 7, 2017
Filed: June 26, 2013
Methods and apparatus related to adjusting the ranking of a suggestion for a geographic location based on determining how many of one or more users visited the geographic location after receiving the suggestion. In some implementations the one or more users may belong to first grouping of users and the ranking may be adjusted only for suggestions to the first grouping of users. In some implementations it may be determined if a second grouping of users visited the geographic location after receiving the suggestion. In some implementations a property of the given geographic location may be determined based on a comparison of how many of the second grouping of users visited the given geographic location after receiving the suggestion and how many of the one or more users visited the given geographic location after receiving the suggestion.
I haven’t been seeing many recommendations of places or events to go to, but the patent offers some suggestions of a wide range of types of activities that might be suggested that sound interesting:
The suggestion system 130 may provide suggestions for a given geographic location to one or more users. The suggestions may be indicative of a potential user activity at the given geographic location. For example, the suggestion may be for a fine dining establishment in downtown, for watching a sporting event at a sporting bar and grille, for a menu item at the sporting bar and grille, for a book reading and/or signing by a best-selling author at a bookstore, for a book sale at the bookstore, for an opera, for a musical performance, for an art exhibit, for a nearby park, for purchase of merchandise, for travel-related promotions, and so forth. In some implementations one or more of the suggestions may be provided at least in part based on rankings that are associated with the suggestions. The ranking of one or more of the suggestions may be determined by the ranking system 140. In some implementations the ranking system 140 may adjust an existing ranking associated with a suggestion. In some implementations the ranking system 140 may initially determine a ranking for a suggestion.
I did write a post about a Google patent this past summer that offered Future Directions for Google Maps that told us about business recommendations showing up in search results that fit with this patent as something that Google could begin to offer. I also just wrote about a Social Where Next patent that provides suggestions of places to go for members of social groups. I could see Google checking to see if people from those groups followed those recommendations.
This patent also tells us that it might personalized recommendations and tailor them to a searcher’s interests:
For example, one or more users may have indicated an interest in sports-related activities, and the suggestions to the one or more users may include suggestions to purchase tickets to upcoming sporting events, suggestions to visit a sporting goods store, suggestions to join a sporting league, and so forth. As another example, one or more users may have purchased tickets to a particular theater in the past, and the suggestions to the one or more users may include suggestions to see the upcoming productions at that theater. As another example, one or more users may have indicated interest in a particular musical group, and suggestions may be provided to the one or more users for venues where the group may be performing in the future. As another example, one or more users may have responded positively to a particular type of suggestion in the past, and suggestions may be provided to the one or more users that are of the same and/or similar suggestion type.
The patent tells us that it may also offer recommendations that may provide more opportunities and suggestions involving specific locations, such as suggesting visiting an art exhibition before dinner at a specific location, tying in a particular interest with a locational search result (for something like a query for “restaurant A”.)
The patent also tells us that it might offer recommendations based upon the physical location of a user:
In some implementations the physical location of the user may be based on locational data described herein. For example, a user’s physical location may be identified as being in the downtown area and suggestion system 130 may provide suggestions for downtown restaurants to the user. Likewise, suggestions for local suburban restaurants may be provided to users that are identified as being in a particular suburban area. As another example, a user in the vicinity of a mall may be provided suggestions to visit one or more stores in the mall. As another example, the user in the vicinity of a mall may be provided with suggestions to participate in one or more ongoing promotional events at the mall. In some implementations the physical location of the user may be based on identified locational data as described herein.
We know that Google is paying more attention to where people go, and how long they spend at those places, as I wrote about in Google Tracking How Busy Places are by Looking at Location Histories
Suggestions offered under the process described in this patent may not be offered until a certain threshold of people have visited particular geographic locations as indicated in a number of possible different ways:
Locational data may be based on one or more actual and/or indicated visits from users such as those discussed herein. For example, locational data may be based on one or more of locational queries, geolocational data from mobile devices and/or other client devices 110, financial transactions at a given geographic location, user indications of visits to a given geographic location (e.g. check-ins), and so forth. In some implementations utilized locational data may optionally be restricted based on one or more aspects of the locational data. For example, locational data may be restricted based on the geographic location, based on one or more suggestions associated with the locational data, and/or one or more groupings of users.
The patent tells us that it might offer more data to increase a “confidence level” that a person might actually travel to a place recommended, such as whether they visit a social network listing for a place or whether or not a search might have searched for the place, or read reviews for it, or in the instance of a restaurant, have read or viewed the menu for the restaurant after a suggestion to visit the restaurant.
The patent talks about different groups of users that might be given geographic locations recommendations
For example, the grouping of users may include users that share a similar age range, such as teenagers, college undergraduates, retirees, and people over forty, couples in their thirties, and singles over fifties. As another example, the grouping of users may include users that are grouped as a certain user type based on one or more shared attributes. As another example, the grouping of users may include one or more users from a common region, such as a zip code, a county, a business district, a city, a township, a municipal area, a state, and/or a country. For example, in some implementations, the common region may include one or more of the eastern, southern, midwestern, southwestern, central, and/or pacific regions of the United States. Also, for example, other designations such as “small town”, “medium-sized town”, “small city”, and “large city” may be used to identify a common region.
Certain groups, such as telling us that a place is “popular with users with children” or “not a family friendly restaurant.” Interestingly, the questions that Google asks people in their local guides program seem to address identifying such groups. But the patent tells us:
In such an instance the restaurant property may be determined to be “not a family friendly restaurant”, “not popular among users with children” and/or “not popular among users in the second grouping of users” due to the predominantly low percentage of users with children determined to have visited the restaurant after receiving the suggestion.
The patent isn’t telling us much at this point about why certain geographic locations might be recommended. Once we start seeing more geographic recommendations from Google, we may be able to identify what types of things might be influencing those suggestions. For instance, there isn’t mention of the use of Schema in the patent, but we know that Schema can influence whether or not events at certain places are noted in search results, so using those could potentially be a positive thing to implement for a businesses website.