Universal Search is Google's grand plan of integrating all kinds of media, be it books, video, images, maps or WebPages. It also involves extends extending Google's search capabilities across different interfaces including television, Radios, Mobiles etc.

Google is not limiting itself to the internet and wants to extend its proven capability in terms of targeted advertising in other interfaces as well.

Let us have a look at some of the most recent patents Google has filed in this regard and understand the opportunity it presents to advertisers

• Providing Advertising in Aerial Imagery (mention dates)
• Local Search and Mapping for Mobile Devices
• Telephone Network Advertising

Providing advertising in Aerial imagery

Advertisers are always on the lookout for new ways to reach to the targeted audience. The media used could be offline (Billboards, Print, Television, Radio, etc) or online (Online Banners, Online Classifieds, Text ads etc). The advantage of providing Online ads is the ability to track user behavior and better targeting. Google has come up with another innovative way to provide space for advertising in this arena: Ads on Maps.
Google provides three possible methods of implementing this

• The search engine receives a request for a map from the user. The search engine then retrieves the requested map and overlays it with at least one advertisement
  
  
• The search engine receives a generic request and in turn retrieves an aerial image with an advertisement overlaid.
  
  
• The search engine receives a request for a place name or a search term and based on the place name or the search term, selects an ad and superimposes it with the aerial image.

Google looks to superimpose the advertisements on rooftops, sides of buildings, parks, tree tops etc.
The factors that could be considered in selecting an advertisement are:

• The locality
• The place name
• The search term or group of search terms
• Information on the user (past clicks on advertisements, where the user likes to see ads on the map etc)

Along with these factors the algorithm might also look at the scaling of the map. So, if you are looking at a map of the United States, the ads presented would be different and if you are looking at a map of Concord, Massachusetts, the ads presented would be different and could be more localized as well.

The algorithm will have a set of candidate advertisements from which it will select the ads for display based on a ranking priority.

The ranking would again be based on the highest bid provided by the advertisers for a click or an impression and the click through rate of the ads in the past.

In another embodiment of the implementation Google may select advertisements based on keywords present within tags on maps. Map service providers such as Google maps and Wiki maps allow for users to tag areas of interest. As such an advertisement may be selected and displayed based on keyword matching and the locality.

In yet another embodiment of the implementation people, organizations or the government can auction off virtual space to advertising networks. Google, when it receives a request retrieves the corresponding map and places predetermined advertisements at the selected spaces.

Telephone Network Advertising

Advertisers for years have been trying to create demographic profiles of users and then target them more appropriately with ads specifically catered to them. It is especially tough to create demographic profiles of mobile device (PDAs, Cell phones etc.) users, television viewers and radio listeners. Google however has been making great strides in this field as well.

This patent from Google could help an advertising system to serve more relevant ads across a telephone network.

The process involved includes:

• Storing a set of called numbers from the user’s mobile
• Using those called numbers to derive user information
• Using the stored information to target specific ads at the user

Every mobile has a call history log which holds called numbers. Google intends to use these numbers and map them to directories and web pages to derive information on the called numbers. For example my call history log holds the number (603) 621-1451 100.

Google retrieves this number and searches it against directories to derive information on the number.

Google realizes that the number belongs to Target Superstore located in Concord, Massachusetts and that it is a department store. This information, it stores in the first information index.

Google now starts forming associations with each attribute value for (603) 621-1451 100. For Concord it could create associations such as Airlines in Concord, Hotels in Concord, and festivals in Concord and so on and so forth.

For department store it could create associations such as Groceries, White Goods, Lifestyle items etc.
For Target it could create associations such as Furniture, Baby Products, Electronics, and Toys etc.

This, in essence becomes the profile of the user. So in the above case the profile would be
Concord Airlines, Concord Hotels, Concord festivals, Groceries, White Goods, Lifestyle Items, Target, Furniture, Baby Products, Electronics, Toys.

Over? Not yet….Google also intends to weigh the call derived information to enable focused targeting.
The weight of the call derived information would be derived based on the following factors:

1. Duration of the Call
2. Cost of the call
3. Number of times(frequency) the number was called
4. Time lapsed since the last call to the same number
5. Call frequency to a type of organization
6. Call frequency to a type of person
7. Number of calls to a business or organization
8. Frequency of calls to the same number
9. Number of calls to the same number

So, taking the Target example again if the search engine finds that a the number and frequency of calls to Target, Concord  is high, Google might assign it a high weight value and would focus on displaying ads catering to ‘Target’ associations.

The advertisement to be displayed would also depend on matching call derived information with the targeted keywords of the advertisement.

From the advertiser’s perspective, he could keep the scope of targeting broad or narrow. The advertiser could state that he wants his ad placed only on weekends; only to users in a particular area and only if certain keywords match. He could alternately state that he wants to target his ads to users of a particular device

So how does the advertising model work?

An ad campaign for a company would involve a starting and an ending date, the budget for the campaign, the targeted keywords, and the areas to target etc.The pricing structure would be defined on a cost per click/impression/selection

The patent mentions that a conversion would occur when a user consummates the transaction related to a previously served ad.

The Conversion can be defined in one of the following methods:

• The user sees and clicks on an ad which then takes him to the advertiser’s web page. He then makes a purchase before leaving the web page
  
  
• The user sees and clicks on an ad which then takes him to the advertiser’s web page. He then makes a purchase from the web page in a predetermined amount of time (eg: 7 days).
  
  
• The user sees and clicks on an ad which then takes him to the advertiser’s web page. He then performs an action: for example filling in contact details or downloading a white paper, signing in for a newsletter etc.