Guest Blog: SpotX Talks Ad Fraud


DataXu isn’t alone in the fight against fraud.  Before we employ a defense in-depth approach on the buy-side, partners on the sell-side implement their own layers of protection.  SpotX is an example of a partner who has taken it a step further and helped DataXu in our guarantee by agreeing to pay back clients for media purchased on their exchange when a fraud rate is over 3%.

The following blog post was guest written by SpotXchange.

At SpotX, we’ve always believed there’s a shared responsibility in combating fraud and that it’s unacceptable to write it off as a cost of doing business in the digital space. Our goal is not only to coach publishers on quality metrics but also provide tools, data and insights to help them effectively manage their inventory. For buyers, we constantly strive to be a trusted platform known for delivering an audience that is as fraud-free as possible.

We use a combination of automated and manual processes to accomplish this. All SpotX publishers are rigorously vetted prior to receiving demand. An evaluation of the publisher’s reputation, content, ad placements takes place and traffic sourcing techniques are closely examined. Once a publisher is live on the platfom, SpotX proactively screens out all bot impressions, allowing buyers to only bid on human users. We then have an in-house brand safety team audit all impressions for domain spoofing, ad injection, viewability, audibility and other impression quality statistics on an ongoing basis.

We believe that there isn’t a simple way or a single technology to eliminate fraud. That’s why SpotX has partnered with several industry leading vendors, utilizes proprietary technology and has a full-time brand safety to combat fraud, eliminate unsafe placements and have a holistic view on inventory quality. SpotX also maintains compliance with industry initiatives around quality, including being certified by the Trustworthy Accountability Group as well as the Digital Trading Standards Group’s Good Practice Principles.