How to Stop Losing Money to Online Ad Fraud
Samuel Scott, Director of Marketing communications at Logz.io, Tel-Aviv-based company producing software that analyzes server log files for SEO, is speaking about online ad fraud and how to avoid it if you are an advertiser.
Understanding the mechanism of online advertising
In the West, $6.3 billion of advertising money is lost every year to ad fraud. It happens because people do not know what an impression is and what its implications are.
Most advertisers think that one impression is one person viewing an ad for one time. This is not true. An impression is a web browser telling an ad network to serve an ad one time. It’s one machine talking to another machine.
So, when a campaign report says it had 100,000 impressions, that was a web browser asking an ad network for an ad 100,000 times. No human beings, only machines are involved – which leaves a door open for a lot of net fraud.
Most impressions are never seen by humans. Exact percentage is unclear – reports say that from 20% (Bloomberg) up to 85% (MediaPost) of your online ad is never seen by real people. It also depends on the country and the ad network.
Why the ads are not seen?
- Ad Network failure. It might load a wrong ad, or not load it at all.
- Bot traffic. Study by The Atlantic said 60% of all internet traffic is bots, and a bot visit still counts as an impression.
- Publisher fraud. Study in Adweek said 25% of the entire online advertising market is fraudulent.
Publishers striving to maximize revenue from ad networks create ad units 1*1 pixels, which are too tiny to see but they still load and count as impression. They also stack multiple ad units on top of each other, so that all of them make impressions, but only the top one is seen. Redirecting visitor to a landing page or a pop-up that visitors do not want to see also counts. Showing ads in website footers or down the page, where person might not even scroll down to, is another wat to deceive advertisers.
Buying bot traffic. Digiday says publishers pay ad networks $0.002 for a bot visit, and ad networks will knowingly sell this traffic. When a bot hits a website, an impression is created and publishers are being paid .0025 to .004 in ad revenue. This makes 25% to 100% profit margin for doing nothing.
Thus, ad network technical failures, bot traffic, and publishers’ fraud make only 8% of online display ads even have the chance to be seen by human beings. The 92% of non-viewable impressions still make money for publishers, ad networks, and media agencies, while the advertisers are being screwed.
What is being done?
- This year in the U.S. only, $25 billion of online ad spending is being under review. This means CMOs are looking at where the money is being spent and think of getting a better ROI elsewhere.
- Google has deployed ActiveView, which is a way to increase viewability so that advertisers pay only when the ads are viewable.
- Facebook is also working on something similar.
- U.S. Interactive Advertising Bureau (IAB) has a goal for the online advertising industry to reach 70% viewability.
How to protect yourself?
- Never run CPM campaigns (cost per thousand) – they are the quickest way to lose money in digital marketing.
- Revise analytics in human terms. The real CPI (cost per impression) is much higher if viewability of 8% instead of 100% is considered. Proceed only if that ROI is acceptable.
- Revise your metrics in CPC campaigns (cost per click). For instance, if your click fraud estimate is 25%, use the same calculations to adjust a CPC campaign to reflect human terms. Do this for all online advertising metrics.
- Run CPA campaigns (cost per action): you only pay when a visitor does something, e.g. buys a product, submits content, etc. It’s a lot harder to do fraud here.
- Block fraudulent traffic countries from your display ad campaigns. The chart below shows current top three locations (the U.S. listed for comparison purposes). Block them from your ad campaigns unless you have to sell there.
You can watch related webinar here