March 30, 2007

Patent for Behavioral Targeted Ads in Games

After acquiring AdScape, a company that produces ads for video games, Google has a new patent that reveals some interesting ways of targeting ads to gamers.

Some context:

"In-game advertising is becoming extremely popular. This trend is expected to continue since the 18 to 34 year old male demographic in the U.S. is watching less TV and spending more time playing video games than ever before. The video game industry is becoming a media force on par with the television and motion picture industries. Consequently, ad agencies and game producers are collaborating to introduce more ads into video games. Presently, in-game ads are used to advertise real products and services in a manner analogous to product placement in movies and television shows. For example, a decal on a virtual race car may advertise a product or service. (...) Unfortunately, ads placed in various video games are typically determined while the game is developed and are therefore relatively static. Further, the ads are typically targeted to a broad demographic group. Consequently, in-game ads are often not as relevant and useful as they could be. "

The patent suggests a system that takes into account user's interaction with the game, its decisions and preferences.
The game play tracking operations may track game player input information. For instance, in most simulation games as in a Formula One racing game, a user may select a real world make/team of a car (e.g., Ferrari, Williams-BMW, McLaren-Mercedes, Renault, etc.), a particular driver (e.g., Michael Schumacher, Fernando Alonso, Rubens Barrichello, etc.) as well as the racing track desired to compete in (e.g., Monte Carlo/Monaco, Nurburgring/Europe, Indianapolis/USA, etc.), car color, type of tires, etc. (...) If a user selected a racing car from Dodge, the system may show a Dodge ad or something related.

Game state-based information may include information about the user's game-play. For example it may include, how fast the players are going through the levels, how familiar the players are with the game, what level are the players in (...), how long have the players been playing the game, how frequently the user plays various games, play-pause habits, game information stored to non-volatile memory, etc. (...) If the user has been playing for over two hours continuously, the system may display ads for pizza-hut, coke, coffee and other related goods.

The play characteristics of users, particularly in online RPG games (e.g., time spent chatting with other virtual players versus fighting, time spent bartering versus stealing, time spent exploring versus building, time spent trying new items versus completing levels, decisions made by players leading to certain situations (good versus bad, strategic versus short term), avoiding conflict (risk averse) versus being aggressive, cooperating and collaborating versus doing things alone, friendly versus hostile, etc.) may be particularly useful. User input information may be useful to help infer information about a user. Some other examples user information that may be inferred includes familiarity with a game(s), time spent playing a game(s), how fast is the user advancing and/or skill level, etc. Inferences drawn from such user input information may be made using known classification means such as neural networks, Bayesian networks, support vector machines, etc. Such inferred user information may be useful to help target ads. For instance, users that spend a long time bartering instead of stealing in a game may suggests that they are interested in the best deals rather than the flashiest items so the system may show ads reflecting value.

Basically, you're living in a virtual world where your behavior is tracked and used to deliver ads that fill some predetermined spots. Behavioral targeted advertising is already used on the web by some ad networks, but the information that is tracked is much more limited (the sites you visit, the length of visit). In games, you have access to subtle details (game's choices could become a personality test). I wonder if gamers would accept such a system.

{via SEO by the Sea.}

2 comments:

  1. A broad envelope patent for targeted ads is held by Concurrent Computer Corporation via its subsidiary Everstream.

    Patent# 6,161,142

    Patent# 6,038,591

    Patent# 5,931,901

    ReplyDelete

Note: Only a member of this blog may post a comment.