Scientists that develop algorithms for automated bidding refer to the process of buying ads on search engines as an “ad auction.” The process of buying ads is really just a continuous auction for advertising space on a search engine. A user that types in a search phrase on Google might get back a list of ads on the top, and to the right, of the search engine. How those listings appear depends on how well the advertiser did in the auction. The winner of the auction is determined by price and other complex variables used by the search engine to help with relevancy and keeping the user of the search engine happy at the same time. There are three participants in the auction that all have different needs: the search engine, the user, and the advertiser. Keeping a balance between the three is crucial if the ecosystem is going to survive.
Most people have heard of automated bidding on the stock market. In that scenario, scientists write programs that look at data about a particular stock and then automatically make trades based on the outcome of the computer’s calculations. The technology uses statistics and machine learning to optimize the buying and selling over a period of time. The more data the programs learn from, the better they become at the auction. Ads on a search engine, like stocks, are a form of an auction. The kind of technology used to automatically buy and sell stocks is also used to buy PPC ads in ad auctions.
Automated bidding algorithms can look at all the historical site analytics, and account data, from your PPC accounts and make decisions based on the historical performance. Humans that make changes to bid prices in PPC accounts are (for the most part) guessing — they can’t possibly keep track of all information there is to know about the users that come to a site and how the auction has performed over a period of time. Automated bidding can find gaps and patterns that humans can’t find — humans can only track a limited set of information while making a change. However, it does still take a good combination of humans plus machines to be profitable at PPC. Automation doesn’t replace the human component of successful advertising, but in the long-term it will beat any agency that only has people making decisions based on instinct and not data.
Sean periodically teaches as an adjunct professor on the topic of search engines and search marketing at MSU and is a member of their computer science advisory board. He completed coursework for his PhD in machine learning at MSU. He was the founder and publisher of SEMJ.org. Sean holds four engineering patents, has a B.S. in physics from the University of Washington in Seattle, and a master’s in electrical engineering from Washington State University. As president and director at metric ppc, Inc. he focuses on search marketing, internet research, and consults for large companies.