Confirmed success with FouAnalytics

The following is a great example of how a global advertiser is using FouAnalytics to monitor their own campaigns and ecommerce site. The screen shots and examples below are taken from FouAnalytics.

Should you be scared of "red"?

If I showed you this on-site chart from FouAnalytics, would you be scared by all that dark red? Maybe ... but,

... if you look at the UTM_SOURCE data grid, it tells you that virtually none of the dark red came from paid media sources. If it did, then the advertiser would have to take action to reduce the fraud and bots, but in this case, most of the bot traffic on the site was just bots directly hitting pages. This is a nuisance but the bots are NOT eating up valuable ad budgets or faking clicks to the site. So no action was needed in this case.

Paid channels were well optimized already

This advertiser buys and manages their own digital media, and they use FouAnalytics to monitor and manage that media. As you can see, their paid channels were well optimized already (mostly dark blue, and very little dark red).

In the data grid, the 10.3% at the top shows the portion of site traffic that has a UTM_SOURCE (that implies 10% of the pageviews came from paid media sources, the rest was direct traffic). This is great, because the advertiser was not buying too much traffic. If I see large percentages of a site traffic from paid media sources, that is a sign of desperation. Not enough humans know about the brand and come to the site directly and the advertiser is just throwing money at advertising to prop up traffic numbers. This is not the case here, since the paid traffic percentage was only 10%.

Let's now look at the breakdowns of the paid sources. The first row is email marketing. This is a great example of how an advertiser "engages" with people who have explicitly given them consent to contact them regularly, not just consented to being bombarded with more ads. The second row shows "google-ads." Again, this is well optimized already because these clicks that arrived on the site were mostly dark blue and very little dark red. "fbclid" means clicks on organic content on Facebook, while "meta" means the paid ads on Facebook that clicked through. All of these are well optimized already with lots of dark blue and little dark red. To connect it to the point above, the dark red (bots) came to the site directly, and those bots don't appear in the UTM_SOURCE data grid because they did not come from paid media sources.

So the overall site might have a lot of bot traffic, because any bot can hit any public website. But the advertiser does not have to take any action because that traffic didn't come from ads or faked clicks on their ads.

How do "google-ads" perform?

Taking a closer look at the second largest source of paid traffic, we isolate just the pageviews coming from "UTM_SOURCE:equals(google-ads)." We see really good characteristics. The donut chart on the left shows 87% dark blue (humans) and almost negligible amount of dark red (2%). More importantly the click charts on the right show that mobile users (portrait click graphs) have about 30 - 37% attentiveness (clicks) while the desktop users (landscape click graph marked 1920x1080) have about 55% attentiveness. This shows that humans on desktop computers are more attentive on the site than humans visiting the site on mobile devices.

How do paid channels compare to direct traffic in attentiveness?

If we now compare the attentiveness of users from various paid channels to "direct traffic" (yellow highlight), we also see that this advertisers' ads even over-index direct traffic in attentiveness. Note the percentages of valid-clicked, mousemove-exists, valid-scrolling, and touch-exists. All of those percentages are higher than the percentages in yellow. That means visitors from paid channels did MORE on the site than users coming directly to the site. Furthermore, there are more confirmed humans (dark blue, right side) coming from clicks on paid ads than direct. Do you know how RARE and awesome this is, having looked at digital campaign analytics for 25 years? It is SO rare that advertising campaigns deliver users that are MORE attentive than the users that came directly to the site. This is why this case study is exceptional, and why we decided to share it. It can be a shining example for other advertisers to strive for.

The way to generate the report above is to run the following in the reports tab of FouAnalytics (only advanced users are given access to this reporting functionality)

What about the ecommerce transactions?

Since this is an ecommerce site and users can buy wine from it, we looked at the "checkout-success" page too. By isolating just the urls that contain checkout-success, we see that most of the conversions were humans (dark blue) and only a handful are dark red (see left donut chart).

Once we isolated the sessions on the site that ended in an ecommerce purchase, we can also look at characteristics of the session. This is neatly summarized in the excerpt of the spreadsheet below. This technique was taken from the How to do cookieless attribution with FouAnalytics article here: https://www.dhirubhai.net/pulse/how-do-cookieless-attribution-fouanalytics-dr-augustine-fou-ykwue

Basically you see sessions grouped by fingerprints (anonymous representations of users). We added the Region and City to confirm it was the same person. The highlighted rows show how many pageviews in the session -- 6, 4, 6, 10, 5, 6, 8 pageviews, respectively in the column marked "total" -- and the number at the bottom of the timestamp column tells you how many minutes the user spent before they completed the session and checked out -- 4.5, 21.7, 4.3, 12.1, 1.0, 14.3, 7.7 minutes, respectively in the column marked "TIMESTAMP." These make sense because humans take a few minutes to check out. Contrast that to botnets buying up all of Taylor Swift's tickets within minutes of release, each bot checkout took mere seconds.

Here's how to pull the data above from the Reports tab in FouAnalytics (again reserved for advanced practitioners)

GEO_REGION_0,GEO_CITY_0,fingerprint,TIMESTAMP

"If advertisers can SEE better, advertisers can DO better"

So what?

I finally believe that the golden age of digital advertising is just dawning now. We've been wandering through the wilderness for the last 10 years, wasting money on bots and fraud and not knowing it. But this shining example shows us the better way forward. It's not that their site had no bots. But with better and more detailed analytics, like FouAnalytics, they can take the appropriate action or understand that no action was needed, despite the large amount of red on the website. Other advertisers can do the same. If they upgrade their tools they can see better. If advertisers can SEE better, advertisers can DO better. Do you want to "see Fou yourself" too?