Online programmatic advertising is a huge, multi-billion dollar per year industry. Ad spend in this area is expected to reach over $300 billion next year. As well as this, other forms of online advertising are expecting similar growth.
The ease of programmatic ad buying and the vast growth of the market has lead to a rise in ad fraud that has infected the advertising ecosystem. Some estimates state that up to $42 billion will be wasted in 2019 due to fraudulent ads.
Some measures have been taken to protect against some basic types of ad fraud. But to counter more sophisticated fraud, advertisers need a more robust solution to counter the huge amount of ad fraud that exists in the industry.
Location could be this solution. Understanding device location and historical behavior can help to identify fraud better than other methods. In this post, we’ll look at the ways that advertisers can use location data to reduce ad fraud and limit the damage from malicious actors in the advertising ecosystem.
What is ad fraud?
Ad fraud is the practice of fraudulently impersonating online advertisement impressions, clicks, conversions or other KPIs in order to generate revenue.
Eliminating ad fraud with location
Does the device exist?
The first step is to identify if the device is a real device and not an emulator. Emulating a device is a common way of faking ad impressions and clicks. In some cases, emulators can even generate IP addresses to pass as a real device.
In this case, location needs to be more precise; the ad inventory needs to combine with sensors to verify that the device exists.
Are the ads landing where intended?
Let’s say that you have a campaign running in the US, and you are only targeting devices in the US. You can use location data to map where the ads are being delivered to the device. If there is a considerable disparity between targeting and delivery, then it’s highly likely that some of your campaign inventory is fraudulent.
Identifying fraudulent inventory is essential as often devices can change or move, and the targeting solution will not update these. But other times, audiences can contain bad inventory, deliberately including devices that don’t meet the criteria. This is why you should always carefully vet your data and audience providers.
Countering smarter ad fraud
What about eliminating the more intelligent fraud? Some ad fraudsters are generating fake IP addresses to spoof IP location monitoring.
Using a powerful location SDK can eliminate this. An SDK uses many signals to identify device location with greater accuracy correctly. Subsequently to trick a location SDK into registering a click in a fake location is much harder to do than a simple IP trick.
What about click farming?
In some cases, fraudsters will pay a real person to click and interact with specific ads in several different locations. Sometimes these devices are kept in one place; other times, they are the person’s personal device.
Location data associated with a device can be used to see if the device moves and behaves like a regular device. Understanding if a device stays in one place and combining this with other fraud detection methods, such as time to install, can dramatically increase the identification of ad fraud.
Towards a version of location ad fraud detection.
Integrating precise location into your ad stack can have a substantial positive impact on your ability to detect ad fraud. But using location data in a more traditional way can also help to understand if clicks and conversions are being bought.
For example, location-based attribution is the process of measuring if an exposed device eventually visits a physical location. This can also act as a way of vetting audiences and inventory.
Location data has many powerful applications across the advertising ecosystem. Detecting ad fraud is another application that allows advertisers and marketers to use location to reduce wasted budgets and identify partners that may be supplying them fraudulent inventory.