FouAnalytics - Frequently Asked Questions FAQs

What is FouAnalytics?

FouAnalytics is an analytics platform for monitoring programmatic ads and also websites. It uses javascript tags which can be copied-and-pasted into programmatic ad campaigns or installed on sites like Google Analytics. It has fraud detection baked in to look for bots (IVT - invalid traffic) and other forms of fraud committed by sites and mobile apps. FouAnalytics provides complete supporting details so you can understand why something is marked as fraudulent or not fraudulent, unlike other black box fraud detection which does not provide any explanations to back up the IVT% they report to you. You cannot troubleshoot campaigns with that and you are forced to keep paying them; with FouAnalytics you can see which sites and apps are cheaters, so you can turn them off by adding them to block lists. This way you can clean your campaigns while they are still running.?

Screen shots of the FouAnalytics dashboard are in this embedded slideshare:

I already have fraud detection, why do I need FouAnalytics?

Commonly used fraud detection tech companies simply fail to detect most forms of fraud. You don’t have to take my word for it; just run a side-by-side comparison with FouAnalytics (no cost to do such a pilot measurement). Fraud detection is tuned to look for invalid traffic (IVT), commonly known as bots. But when we are talking about mobile apps, connectedTV, etc. these forms of programmatic ads don’t involve bots going to webpages. So those technologies fail to detect most of the ad fraud, especially considering that mobile ad spending represents the vast majority of display and video ad impressions right now. Also, these technologies don’t even look for shady activity from the sites themselves, like ad stacking, pixel stuffing, popunders and popups, forced redirects, auto-refreshing ad slots, rotating slideshows, etc. FouAnalytics looks for all forms of fake users (bots) and all forms of fraud on sites and mobile apps and the data shows you why something is considered fraudulent so you can decide whether to continue buying from those sites and apps.?

For more details of how badly fraud detection tech fails, see: https://www.forbes.com/sites/augustinefou/2020/11/23/what-your-fraud-detection-vendor-misses/

https://www.dhirubhai.net/pulse/why-black-box-detection-tech-sucks-fou-ad-fraud-researcher

The exchanges I buy from all have prebid filtering, why do I need FouAnalytics?

At the time of the bid request, before any bids are submitted, every parameter in the bid request is declared, and thus every parameter can be faked. You should always assume it is fake until proven real. Let me explain. If a fraudster had a site fakesite123.com would they ever put that domain in the bid request? No; they would never get any bids. So fake sites will always pretend to be some other domain -- e.g. marthastewart.com -- this is domain spoofing. Prebid filtering catches some of these tricks but not all the time. In a more common scenario, a less obvious fraudulent site like coolcelebritypics[.]com might actually use their own domain in the bid request. If prebid filtering has seen a known fraud site before, it will avoid bidding on it. Once a domain stops making money, the fraudster rotates the domain out and starts using different ones, thus defeating this form of protection against fraud. The same occurs for cookies. If a cookie is known to be a bot, prebid filtering will not bid on it. But the moment a cookie fails to make money for the fraudster, they dump the cookie and get a new one, again defeating prebid filtering. FouAnalytics runs after the ad is served so we can see where the ad actually ended up -- which site or mobile app. If these are known cheaters, we add them to the block list. All the fraud that gets through prebid filtering can be seen this way and removed by FouAnalytics.

Having continuous monitoring with FouAnalytics is better than prebid filtering and fraud detection, see: https://www.forbes.com/sites/augustinefou/2021/01/19/analytics-are-better-than-fraud-detection-heres-why-with-examples/

Does FouAnalytics do real time blocking?

No. And we will never do real time blocking. Real time blocking assumes the detection is always correct. And as we saw above, the current fraud detection tech is very inaccurate -- it doesn’t catch the fraud that is there, and it labels non-fraudulent things as IVT and blocks it. Case in point, a fraud detection vendor marked mainstream sites like esquire[.]com, travelandleisure[.]com and other sites as high IVT and blocked them. This was due to incorrect measurement.

The real sites got blocked incorrectly, and the fraudsters' sites were not blocked. This is an example of the consequences of allowing algorithms to not only detect, but also take action like block the domains. In a case where FouAnalytics was used, a particular domain was recommended for blocking; but upon human review the cosmetics company had a direct relationship with the publisher and wanted to continue to work with that domain; by disallowing the ads on certain portions of the sites were purchased traffic was used, the client could reduce the fraud in their campaigns while continuing to work with the publisher. The business decision was left up to the client, rather than a black box algorithm automatically blocking the domain, without providing any reason or supporting details.?

See: Examples Of Incorrect Measurements By “Black Box” Fraud Detection

Is FouAnalytics MRC accredited?

No. I will never, ever, seek MRC accreditation or TAG certification. This is because the MRC and TAG are not qualified to accredit my tech. They have no tech, no data set, and no “answer key” to know if anything is measured correctly. TAG “certification” is obtained by completing paperwork and paying the annual fee -- i.e. “self-attested.” Even the ones that are marked “independent validation” the validation is done by another non-profit that also has no tech, no data, and no answer key. They are reviewing the completeness of the paperwork and the existence of the company behind it, not the accuracy of the detection of fraud. MRC accredits companies for doing what they said they would do. So a company that said it would measure 3 things will get accredited just like another company that said it would measure 300 things. In neither case, can MRC tell if they measured fraud correctly, just that they were measuring what they said they would measure.?

In the case of FouAnalytics, all the supporting data is there for you to review yourself. I will explain what things mean and if I can explain it to you in a way that you understand and agree with my assessment, you can take the remediation actions yourself -- e.g. add those sites and apps to your block lists.?

See: Do MRC and TAG certifications help reduce ad fraud? https://www.forbes.com/sites/augustinefou/2020/06/15/do-certifications-and-accreditations-help-reduce-ad-fraud/

Does FouAnalytics do brand safety detection?

Yes. We detect the exact url/page and domain on which your ads were loaded. With this detailed log of sites and apps, you can decide which sites and apps to continue advertising on. If there are sites that are clearly not-on-brand for you, you can remove them from your programmatic buys. Just like with real time blocking for IVT above, FouAnalytics does NOT do real time blocking for IVT or for brand safety. FouAnalytics is an analytics platform which records where you ads went so you can make the business decision yourself whether to continue to buy from those sites and apps, and add to your block lists (or remove from your whitelists) as you see fit. The platform does not make those business decisions for you automatically.?

Is FouAnalytics privacy compliant?

Yes. FouAnalytics is privacy regulation compliant; it collects only anonymous javascript parameters from the browser; no PII (personally identifiable information) is ever collected; no 3rd party cookies are ever set or used. I am an independent researcher and not an ad tech company. The data collected by FouAnalytics has no other use, other than to show clients whether any fraud impacted their campaigns. Further, in the EU, we have dedicated server stack, so data never leaves the EU; also since the strictest interpretation of GDPR is the German interpretation which regards IP addresses to be PII, we discard the IP address as well for all EU visitors.?

Is FouAnalytics secure and tamper-proof?

Yes. FouAnalytics has cybersecurity measures built into every part of the system so false data cannot be written into it; analytics cannot be altered or tampered-with, unlike Google Analytics were false data can be easily written into it. (exception - those sites that have upgraded to GA4 are less vulnerable to tampering.)

Cybersecurity measures built into FouAnalytics: https://www.dhirubhai.net/pulse/does-your-analytics-platform-have-cybersecurity-dr-augustine-fou-/

Loopholes in Google Analytics Puts Marketers Insights at Risk: https://www.forbes.com/sites/augustinefou/2021/04/22/google-analytics-vulnerabilities-put-marketers-at-risk/

Please let me know if you have other questions. Either email me or post those questions in the comments below and I will answer.