Fake News Algorithms Cause Loss of Revenue Among Advertisers

This past weekend was a fun excursion into the realm of marketing and advertising. Specifically, how a marketing campaign can be potentially wasted because of poorly written code.

Photo by Fabian Grohs on Unsplash

A client last weekend came to me with a problem: for whatever reason, an ad would not populate correctly on his website.

At first, I thought it might be a problem with the ad agency, so I had reached out to them; no response. Then, my client was able to get an email back which, in summation, said this:

We don’t know what happened, it’s probably your fault.

Cheers!

-Management

So I thought, maybe it is our fault. Which led me to ripping apart the problem page by sections, analyzing those sections, and then piecing it back together.

Whenever I face a problem, I always pull from my hat that trick that your Math professors probably do as well. The one that goes: “Choose a number between 1 and 100, and in 6 guesses or less I will have chosen that number!”

I translate that into my business. I will take the entire page, network, website, etc. and distill it into a range of possibilities. Some potential errors are encapsulated by others, so I group them accordingly until we have the entire Object. Once the object and the groups are identified, I pick them apart. One by one, until I find the issue.

The issue was, actually, an issue with the filtering software used by the Ad agency. The ad agency had bought a code block from a certain well known company, and had embedded that code-block into their JavaScript for use in filtering out “Fake News” Websites. Those filtering algorithms were producing a lot of false positives, however. Every website similar to my client’s, who also had tried to show the ad, had the ad blocked by this filtering mechanism.

I found the error was linked to “Fake News” filtering. Apparently the algorithm was looking for company names embedded within permalinks, and automatically blocking the ads when the permalink showed, for example, “Ben-and-Jerrys”. Ironically enough, if Ben and Jerry’s wanted to market a new ice cream flavor using Blogs, undoubtedly they’d experience a high loss of conversions and impressions by using this filtering mechanism.

The problem with the filtering was that it blocked all of the websites who were promoting a legitimate product with the name of the product in the permalink. The data analyst inside me was writhing in agony. If your main advertisers are bloggers with a large following, why allow this oversight to ruin a marketing campaign? Would you not logically assume they aren’t changing their permalink structure, and just clicking Publish, considering almost all bloggers use WordPress for its quick posting abilities? If not, then count on a 100% False Positive rate. Which, in the websites I sampled, was the case in this instance.

This weird circumstance affected my client directly. It cost him valuable time and money trying to troubleshoot the issue. Not only did it cost him time and money, but it also made him lose out on potential conversions as well. Oof!

Moral of the Story

If you’re going to try and target wide sweeping key terms for “fake news”, try to check your algorithms first. It could cost you money in the short and long term. And for digital ad agencies, please make sure you test the filtering settings of certain ad filtering programs. You could lose business as a result of this lack of oversight.

Looking for someone to design a website that’s clean, crisp, and easy to manage? I’m your guy! Contact me today! Don’t need a website but no someone who does? Let us know! I pay you directly on referral.

Love the content? Give us a share!
0

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.