Google’s friend-to-friend ad network

If Google has its way, web advertising is about to get a whole lot friendlier – and the definition of spam a whole lot fuzzier.

Recently, the search giant has, as noted by Bill Slawski, filed a series of patent applications that point to a new, more personal way of targeting and delivering advertising. Unlike Google’s highly automated AdWords and AdSense systems, the method relies heavily on manual labor and individual judgment. Rather than being selected by software algorithms, ads would be chosen by people for insertion into their emails, instant messages, and other personal communications. The proposed system puts a crowdsourcing twist on ad targeting. More darkly, it seems intended to put a gloss of respectability, or at least friendliness, on what previously might well have been viewed as spam.

Th system has three essential components. In a patent application titled “User Distributed Search Results,” Google researchers describe the concept of “a universal distributed search system [that] allows users to find and distribute search results (possibly including advertisements) to those with whom they communicate. The search results can be easily distributed by the user via a simple interface that allows the search results to be easily added to the user’s content.” Google has seen that people often put links or other references to related content into emails and other messages they send to friends and acquaintances. The tool described in the patent makes the discovery and inclusion of such content simpler and faster. In essence, it establishes a new way to syndicate Google search results, from friend to friend.

In a second application, titled “Facilitating manual user selection of one or more ads for insertion into a document to be made available to another user or users,” Google researchers describe an associated tool for allowing individuals to insert into their messages ads related to the search results they include. The ads can be either automatically generated, as in the AdSense system, or chosen individually. The tool, as the researchers describe it, “facilitates insertion of manually selected ads into a document that is to be distributed (e.g., transmitted, published, and/or posted) such that the document is to be made available to other users. For example, manually selected ads can be inserted into an email to be sent to another user, a blog to be posted for viewing by other users, a message to be sent to another user, a message board entry to be posted for viewing by other users, a document published and made available to other users, etc. Hence, [user-distributed] ads provide a scaleable advertising platform that achieves at least some of the benefits of manual targeting.”

The patent filing describes a number of ways the system might work, including this one:

Consider, for example, a user that sends an email to members of her book club informing the members of what next month’s book is. Suppose that the user has manually inserted into the email “results” such as an image of the book cover, a UDS search result to a review of the book, and a normal amazon.com search result. When the recipients of this email open it, side-bar, content-relevant ads might also be provided. Such side-bar, content-relevant ads might have been automatically determined using, perhaps among other things (e.g., the textual content of the email message), information derived from the manually inserted “results.” For instance, Amazon might have an ad offering free shipping for purchases made in the next 48 hours.

The final component is a reward system to provide people with incentives to include ads in their personal communications. This component is described in a third patent application titled “Providing rewards for manual user insertion of one or more ads into a document to be made available to another user or users, for distribution of such documents, and/or for user actions on such distributed ads.” Such rewards, according to the filing, “might include one or more of (a) a monetary amount, (b) an enhanced reputation or reputation increase of [the user], and (c) a credit.”

Google’s automated ad delivery system has been enormously successful, generating huge revenues and profits for the company. But the automatic targeting of ads remains imperfect. As the Google researchers note in their filings, “Although advertising systems such as AdWords and AdSense have proven to be very effective tools for advertisers to reach a receptive audience, even automated systems that use sophisticated targeting techniques often can’t match the effectiveness of manual targeting.” The new system that Google has under development would help solve the problem by incorporating the judgments and the relationships of individual people into the selection and distribution of ads. At the same time, it would also mean the injection of commerce, and commercial incentives, into the most intimate of electronic communications.

4 thoughts on “Google’s friend-to-friend ad network

  1. AndyBeard

    This is just a spin of CPA and affiliate marketing, and I can’t see that much that really warrants a patent.

    Even a rewards scheme is similar to various coupon and reviews sites.

  2. Mike of Concrete

    Interesting idea, but right now, a friend-to-friend e-mail may look something like, “Hey, do a Google search on {these keywords} and check out the video,” or else a link to the search itself, copied out of the browser.

    Google presumes that someone will want to do some extra work to inject ads into that “hey” e-mail. They can patent this, but they can’t make people use it unless they come up with a compelling reward for spamming your friends. A couple of pennies isn’t going to cut it here.

  3. Sergey Schetinin

    Back in May I wrote a short blog entry predicting something like this would happen.:

    “Oh, wait, this could be taken yet further, there should be an AdSense kind of thing around this. Anyone would be able to subscribe to the service, install a small application on his cellphone and then get paid every time he mentions certain products in appropriate contexts (the cellphone app is required to track the conversations, and not just over-the-phone ones). Some people would cheat at this just like with click-fraud, but whatever.”

    from Joys of Personalization

Comments are closed.