When it comes to Ad Fraud Detection and Prevention, Pre Bid and Post Bid are where most companies do their logic and analysis.
Compare it to owning a physical home.
Digital advertising transactions occur in the blink of an eye, literally, the average transaction occurs between 1/10th to 3/10th of a second. In that window, an entire auction takes place with multiple technologies communicating with each other which occurs in three main stages Pre Bid, Bid and Post Bid, the latter is after the transaction completes.
Pre Bid is the very start of the whole process and occurs when a user lands on a property like a website or an app and the ad unit begins to render and auction off their advertisements. Those requests are sent to external buyers with detailed information regarding the user, the property, and the device.
Bid happens when the potential buyers receive the data and make a decision on it resulting in them choosing to either bid or to pass. If a winning bid is placed, it is returned back to the property who requested it and the advertisement is rendered to the user.
Post Bid occurs after the advertisement has been rendered, or the auction process has been completed in the case of an unsuccessful bid. Any changes or analysis of the data following such an event is all considered Post Bid.
When it comes to Ad Fraud Prevention, Pre Bid and Post Bid is where most companies do their logic and analysis. Yet the two are not created equal, Pre Bid is typically a query lookup to a pre-populated list or database of data to see if the user on the website or app was flagged by the Post Bid analysis in the past. In other words, it is a query to previous Post Bid analysis which is old data.
This means all the high-powered artificial intelligence is happening in Post Bid, which once you understand the limitations of AI makes sense as you do not have any time restraints in the Post Bid stage. Since AI analysis and human analysis take far longer than “the blink of an eye” on average, Post Bid is the only stage available to do it.
What this equates to is fraud prevention that doesn’t actually prevent fraud but merely identifies potential fraud based on historical data. To illustrate this better we can compare this to owning a physical home.
As a homeowner, you purchase insurance to protect it against financial loss in the event of certain events. Now imagine a day comes when you are robbed, you call the insurance company and they write you a check, and financially you are made whole. However, you are robbed again, then again, and again. The insurance company is writing you checks each time but you are continually being robbed, you complain to your insurance company and they claim to have a solution: they install locks on your doors, with one caveat you have to keep the key under the mat outside.
After all of this, you are still getting robbed, and they are still writing checks and the locks remain.
This is the state of advertising fraud today. There is very little prevention, only insurance,
Pre Bid detection is lacking and in most cases no better than locking a door with a key under the mat. In the end, your house doesn’t just need insurance, it needs a security system, something designed not to write checks but to keep you safe from fraud in the first place.