Save Time, Save Work with FouAnalytics

Advertisers transitioning from legacy verification vendors to FouAnalytics were surprised when I told them they could save significant time and work by setting a single verification tag at the campaign level; i.e. they DIDN'T need to tag each ad creative with a different verification tag. I was surprised that they were surprised about that.

But think about this for a second. The spreadsheet below (from a legacy vendor) shows the invalid traffic rate by ad creative -- e.g. 320x50c-s, 300x250a, etc. What action would you take with this data if you saw that creative A had 0.5% invalid traffic versus creative B with 0.8% IVT? Right, nothing. It doesn't matter that one creative had slightly more or slightly less IVT. Bots don't care what creative it is; their job is to load all ads to get paid.

Advertisers can save significant time by setting one verification tag at the campaign level rather than different tags in different ad creatives. The reason agency folks are so overworked is because they are doing so much unnecessary work setting different tags on different creatives; but agencies are not complaining because all of that is billable hours that you, the advertiser, pays for. The enormous amount of data generated is also unnecessary because it is not actionable. This is what held advertisers back when they were using legacy verification vendors. Not only did the vendors miss catching most of the fraud, they also caused a lot of unnecessary work, that resulted in spreadsheets with lots of data that was not actionable anyway.

I learned from some advertisers considering FouAnalytics that they were concerned about having to do a lot more work compared to when they used legacy verification vendors. Yes, there is more work, in the sense that you have to be willing to look at analytics yourself. But it's not a lot of work, and the learning curve is not steep. Let me explain. Over the years, I have optimized the FouAnalytics dashboard to help me get to the action items as quickly as possible. For most programmatic campaigns, that meant finding the most egregious bad guys and adding them to the block list. That will have the greatest impact on cleaning up the campaigns.

While there may be 100s of thousands of sites and apps in any programmatic campaign, we don't need to look through thousands of bad sites and apps. That's because a site or app that is only eating up a few dozen impressions doesn't matter that much, compared to bad sites and apps that are eating up significant percentages of your impressions. In the Domain App report tab in FouAnalytics, we surface the top 10 "worst offenders" for you to review. In the screen shot below, note the column called "prevalence." The percentage means the percent of your total impressions eaten up by the site or app that is flagged. The first site below is eating up 23% of the impressions (almost a quarter); the second site is eating up 7% of the impressions.

Reviewing these top sites and apps for the various forms of fraud (columns to the right) allow you to decide whether to keep them or block them. Note the red color coding under each box. That tells you the percentage of impressions that got flagged for that form of fraud. If you look rightward, across each row you will note the percentage and the types of fraud that the particular site or app got flagged for. If the site or app got flagged for 3 or more types of fraud, you should seriously consider blocking it. You do so by checking the checkbox. This adds the site or app to the lists at the top. You can copy and paste these into your DSP block list.

We recommend you put a process in place to review this data once a month or once a week. See the top 10 - 20 bad guys and consider adding them to your block list. We don't recommend you continuously add sites to the block list. Otherwise, you can't see the changes, like the examples above -- reduced dark red from 55% to 25% and from 48% to 23% by adding bad sites and apps to block list. Take these optimization actions periodically, so you can see the changes. Not only can you optimize away from fraud (reduce the dark red) you can also optimize towards humans (increase the dark blue).

Happy fraud hunting with FouAnalytics... save yourself time and work, with FouAnalytics too.

Further reading: https://www.dhirubhai.net/in/augustinefou/recent-activity/newsletter/