Priorities of different ADVERTISERS using FouAnalytics

Different advertisers are at different stages of maturity when it comes to digital advertising. You might be wondering how that is possible, because we're nearly 30 years into digital advertising. Yep, I've witnessed this evolution first-hand since 1995. What I mean is that advertisers that have been using legacy fraud verification vendors for the last decade have been misled to believe that only 1% IVT ("invalid traffic") and fraud affected their campaigns. They didn't know and couldn't see that ad fraud and bots were dramatically higher, because the tools they used simply couldn't detect it and report it to them. Advertisers that have upgraded to FouAnalytics can SEE better so they can DO better digital marketing now.

This is not just me flexing. Let me show you how different advertisers are using FouAnalytics in different ways, depending on their "stage of maturity."

Advertisers still using block lists (not using inclusion lists yet)

Advertisers who are still using block lists in their campaigns are using FouAnalytics to identify obviously fraudulent sites and apps to add to their block lists. Some advertisers can cut the dark red (bots and fraud) from their campaigns by half or more. Note that while legacy fraud verification vendors continue to tell them fraud is 1%, you can see in the 3 examples below that dark ref (ad fraud) is 55X, 48X, and 43X higher than that. Once you can see where the fraud is coming from (see comical list of app names below), you can add these to the block list. Just note that blocking bad guys is an endless and thankless job because there's an endless stream of bad sites and apps to try to block.

Note the bad apps seen in the campaigns. There will be a constant flood of these if you don't use inclusion lists. There's just too many bad guys' sites and apps to block.

Advertisers using inclusion lists and PMPs

Advertisers who have realized that blocking bad guys is and will never be good enough, have moved on to inclusion lists. But, you can imagine, some inclusion lists have 100,000 different sites and apps in them, while others have 1,500 sources in them. So advertisers using inclusion lists still need FouAnalytics to measure the quality of the sites and mobile apps so they can prune the lists further.

MFA sites - in FouAnalytics, we categorize MFA sites and MFA apps as those that attempt to load too many ads per page, or do other shady things like load ads continuously in the background, in the overnight hours, in 0x0 pixel windows, stacked on top of each other, refreshed every 1 second, etc. FouAnalytics does not block MFA sites; the platform just surfaces the reasons and supporting data so you can see why they are "MFA." The data tables below just show a simple example -- who thinks having164 iframes on a page is normal? How about 20 ads per page? In any case, some MFA sites have human audiences (think click-bait-ey, funny meme sites). The advertiser can use the data from FouAnalytics to decide whether to block these MFA sites or not. There are other outright fraudulent MFA sites with no human audiences, and that use every fraudulent technique under the sun to generate fake ad impressions to sell. Those are the ones you block or remove from your inclusion lists.

Leakage - the other problem that still exists even if advertisers use inclusion lists is "leakage." That means fake sites and apps still get into your campaigns when you don't want them. How does this happen? Fake sites and apps have to lie about the domain or app name in the bid request, otherwise they would get no bids. So a fake site pretends to be a well-recognized site in the bid request; MFA sites do this too. That helps them get around block lists. And if they pretend to be a site or app that IS in your inclusion lists, they are still able to steal your money. Log level data and placement reports don't show you this because they record the domain or app name passed in the bid request, not the domain or app where the ad actually ran. If you run FouAnalytics in-ad tags, we DETECT where your ad actually went, and we can use that data to check that your ads are actually going to the sites and apps in your inclusion list.

Just yesterday, I reviewed data with a client and we saw that 20% of their ads didn't go to the sites and apps in their inclusion list. We tracked down which supply paths the "leakage" came from and turned those supply sources off. So advertisers that are already on inclusion lists, are still using FouAnalytics to ensure their ads are going where they should be, and to further prune and improve those lists. An agency group is also using FouAnalytics to see the relative quality of each domain to adjust their bids per domain.

Advertisers that have minimized fraud already

Finally, those advertisers who are using inclusion lists and who have already minimized ad fraud in their campaigns, are using FouAnalytics to do more advanced optimizations. In the following example, the dark red is sub-10%. It will never be zero because even on mainstream publisher sites, there are some bots that the publishers and the DSPs can't filter out. And the advertiser wants to place ads on those sites and apps. The the dark red is as low as it can go. Note that the dark blue is also very good.

In this next example, which I wrote about yesterday, the advertiser's in-house team was progressively optimizing for more dark blue (humans). This is a video campaign running in North America. Note the stair-step increases in dark blue. When ads are shown to humans, they have an opportunity to work.

Once the advertiser has minimized the red and maximized the blue in the in-ad measurement, they can look at FouAnalytics on-site measurement on their landing pages. This is where they can start to optimize for attentiveness. A quick reminder that "attentiveness" is not viewability or attention, but INCLUDES both of those concepts already. That's because a human who was attentive on the landing page would have had to see the ad (viewability) and looked at it (attention) to be inspired by the creative message to click through.

Once they arrive on the landing page, they will do something like move the mouse, click something, scroll the page, touch the screen (smartphone), etc. The left side of the chart shows "high attentiveness" where 66 - 69% of the users clicked something on the landing page. It makes sense because these users came from UTM_SOURCE=google (paid search). The users on the right side above were LOW attentiveness users that came from programmatic ads. They clicked through but only 7% of those users did something on the landing page. These visitors are obviously less valuable than the high attentiveness ones.

So what?

Hopefully the above shows you some concrete examples of how advertisers are using FouAnalytics to do better digital marketing. The legacy fraud verification vendors that have been reporting 1% fraud for the last 8 years have clearly not been helpful. You might be thinking, those vendors' tech was never designed to do the above. You're right. They were platforms meant to detect bots, originally. Then they added other "features" to upsell to their own customers. You be the judge and consider whether any of those expensive features enabled you to do better digital marketing.

You are welcome to use FouAnalytics to "see Fou yourself" how much more you can do to make your digital campaigns better.

For screen shots and case examples from FouAnalytics, subscribe for more -- https://www.dhirubhai.net/in/augustinefou/recent-activity/newsletter/