From Wikipedia:

In the bid-based model, the advertiser signs a contract that allows them to compete against other advertisers in a private auction hosted by a publisher or, more commonly, an advertising network. Each advertiser informs the host of the maximum amount that he or she is willing to pay for a given ad spot (often based on a keyword), usually using online tools to do so. The auction plays out in an automated fashion every time a visitor triggers the ad spot.

When the ad spot is part of a search engine results page (SERP), the automated auction takes place whenever a search for the keyword that is being bid upon occurs. All bids for the keyword that target the searcher’s geo-location, the day and time of the search, etc. are then compared and the winner determined. In situations where there are multiple ad spots, a common occurrence on SERPs, there can be multiple winners whose positions on the page are influenced by the amount each has bid. The ad with the highest bid generally shows up first, though additional factors such as ad quality and relevance can sometimes come into play (see Quality Score).

In addition to ad spots on SERPs, the major advertising networks allow for contextual ads to be placed on the properties of 3rd-parties with whom they have partnered. These publishers sign up to host ads on behalf of the network. In return, they receive a portion of the ad revenue that the network generates, which can be anywhere from 50% to over 80% of the gross revenue paid by advertisers. These properties are often referred to as a content network and the ads on them as contextual ads because the ad spots are associated with keywords based on the context of the page on which they are found. In general, ads on content networks have a much lower click-through rate (CTR) and conversion rate (CR) than ads found on SERPs and consequently are less highly valued. Content network properties can include websites, newsletters, and e-mails.

Advertisers pay for each click they receive, with the actual amount paid based on the amount bid. It is common practice amongst auction hosts to charge a winning bidder just slightly more (e.g. one penny) than the next highest bidder or the actual amount bid, whichever is lower. This avoids situations where bidders are constantly adjusting their bids by very small amounts to see if they can still win the auction while paying just a little bit less per click.

To maximize success and achieve scale, automated bid management systems can be deployed. These systems can be used directly by the advertiser, though they are more commonly used by advertising agencies that offer PPC bid management as a service. These tools generally allow for bid management at scale, with thousands or even millions of PPC bids controlled by a highly automated system. The system generally sets each bid based on the goal that has been set for it, such as maximize profit, maximize traffic at breakeven, and so forth. The system is usually tied into the advertiser’s website and fed the results of each click, which then allows it to set bids. The effectiveness of these systems is directly related to the quality and quantity of the performance data that they have to work with – low-traffic ads can lead to a scarcity of data problem that renders many bid management tools useless at worst, or inefficient at best.