Using “Honeypot” Banners to Detect Click Fraud

In our previous article, we discussed the prolific damage fake traffic continues to cause the online advertising industry. To recap: approximately a third of the total traffic transmitted through online ad agencies is fake, accounting for over 15 Billion US$ worth of traffic.

Fake traffic and bots have become an accepted reality for advertisers. Most advertisers set a threshold to account for bots, known as “Maximum Level of Suspicious Impressions.” Said threshold for mainstream traffic is typically set to a maximum of 5%.

However, as online advertising continues to grow and evolve, bots are also becoming more advanced. In the past, bots were much easier to detect: by looking for abnormal behavior generated by a common IP, a common user agent, or alternatively, could be easily tracked by source, time on site, bounce rate, number of new sessions, average pages per visit and other behaviors. Advancements in scamming have generated bots with the startling ability to mimic human reaction. Bots can now open several pages and click on many buttons, simulating a human user. These new bots are much smarter and harder to track.

So what is the best method for an advertiser or a domain owner to identify click fraud? At TrafficTraffickers.com, alongside the traditional scouting techniques, we use the “honeypot approach.” It can be your best solution when CAPTCHA is not an option.

When we suspect fraudulent clicks, we create a banner that is not visible to human users. We design it to fit the context of the web properties that we want to control and give it similar nomenclature to the rest of the banners. Basically, these banners are identical to a normal, visible banner, with one important exception: bots will click on them, but humans will not (with a small margin of error for accidental human clicks.)

This little trick is a great way to find your usual suspects, and to cut down on bots and fake traffic.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>