It’s in LinkedIn’s (and other ad revenue driven networks) economic interest to waste your money, but what’s worse is that most in-house teams and agencies are incentivised to go along with this too, as that’s what they’re measured on. It becomes a perpetual charade of fraud and companies wasting their marketing budget.
In general, audience networks are full of click fraud, as the website owners can use bots and anonymous residential proxies to generate fake clicks on the ads. For each of these fake clicks, the advertiser pays money to the ad network, and the ad network shares the money with the website owner.
LinkedIn is owned by Microsoft. Microsoft Ads has less than ideal click fraud detection capabilities. As an example, Microsoft Ads cannot currently detect puppeteer-extra-plugin-stealth. That means if a website owner, running ads being served by Microsoft, uses puppeteer-extra-plugin-stealth, and routes the bot through a proxy service like Bright Data, the fake clicks on their website will be considered real. Multiple website publishers using puppeteer-extra-plugin-stealth to steal millions of dollars from advertisers every year.
When monitoring ads you either need to stay on top of your placement exclusions list (whenever your ads get click fraud, add the scam websites to your placement exclusions list so they can no longer display your ads), or start with every website blocked, and only whitelist the websites you want your ads to appear on.
You must use a good click fraud detection service so you can quantify how much fraud you’re getting, see which websites you need to block, and have details of every fake click so you can apply for refunds.
If you advertise online, consider limiting it to Google Search only, as it has the lowest amount of click fraud due to there being no financial incentive for criminals to click on ads within search, and Google makes an effort to keep bots out of its search results.
Click fraud is a crime in various places around the world, although it’s rarely punished. A lot of it happens in different jurisdictions (e.g. Russian fraudsters stealing from US advertisers) which makes prosecution difficult.
The people doing it range from organized crime gangs to Nasdaq listed publishers.
It’s quite common for publishers to mix in bot traffic with their legitimate traffic. This helps them guarantee a profit, and gives them plausible deniability as the fraud looks like it’s coming from their paid traffic source.
If you want to find out which networks have people activity undermining the network with click fraud just search “[AD NETWORK] earning method.” You will find blog posts before detailing exactly how they are sending the fake clicks.
Another way is to check affiliate forums and see which ad networks have people complaining about not being paid and their account being shut down. 9/10 times these people were sending fraudulent clicks and the other affiliates know it and call them out.
This kind of fraud goes well beyond audience networks. The big money is in search ads because of relevance. Advertisers pay top dollar to appear for X term. Since search is a finite resource and greed is a natural human emotion, we get a lot of gray area and some outright fraud, and that’s not even audience network. That is supposedly “search”
If you want to see a real world example of a large company artificially driving cold traffic to lucrative search terms, look no further than Yahoo’s home page. Look in the upper right section above the fold and see the “Trending Now” links…nice. You get some human/celebrity interest and then “mortgage rates” and “life insurance”
The point is this… You can barely even trust search to be organic. Audience networks are a collection of 3rd party properties that have all of the incentive to inflate revenue by any means they can get away with.
Breaking down ad fraud in search works likes this:
- The publisher creates a website which can display search results. For example, if a visitor searches for “luxury handbags” on the website, he will see search results relating to luxury handbags.
- The publisher contacts an ad network like Microsoft Ads, and opens a publisher ad account. This allows him to display context relevant ads on his website. That means the search for “luxury handbags” will display ads relating to luxury handbags above the search results. If a visitor clicks on one of the ads, the publisher will receive a fee from the ad network. For example, if a visitor clicks on an ad for “Chanel Handbags”, the advertiser (in this case, Chanel) might pay $20 to Microsoft Ads, and Microsoft Ads would share this money with the publisher.
- The publisher researches which types of searches will display high paying adverts. For example, if a lawyer in New York was looking for people to join a class action lawsuit, he might be willing to pay $500 per click. If the publisher can generate a search to display this advert (“New York class action lawsuit”), he will get paid a few hundred dollars every time a visitor clicks on the ad.
- The publisher hires a bot developer to create a bot which will go to his website, perform a search, and then click on one of the ads. There are many ways to create bots, but commonly it’s puppeteer-extra-plugin-stealth routed through a proxy service like Bright Data. The bot will click on the ads roughly 5% – 10% of the time, to keep the CTR within a believable range.
- A problem the publisher faces is none of his clicks convert. Therefore he has to do conversion fraud to trick the ad networks into believing his traffic is real. This consists of clicks (usually manual) on the ads on the publishers websites, followed by an action such as adding an item to a cart, or signing up to a mailing list, or filling in a leads form. These conversions are garbage, but they’re good enough to make his traffic look real.
A year ago I did a deep dive and saw there was a large percentage of traffic coming from Linux users from 2 different IP addresses and they all had an active session time of 2 seconds or less. They were all assigned a different user ID despite visiting the site within minutes of each other.
Another client who targets insurance brokers and they like to run engagement campaigns to get more likes on their page. For over a year now we see profiles liking the ads that don’t have any picture or job description, and other profiles that have nothing to do with insurance like “bus driver”, “retired” or “stay at home mom”. LinkedIn and their support overall leaves a lot to be desired.
In these cases 44% of LinkedIn traffic coming from Linux users while Facebook was at 0.21% for the same campaign. LinkedIn also had 71% of the sessions with an active session time of 0 seconds while Facebook was at 10%.
LinkedIn advertising in general is a waste of time. It seems like the perfect setup—every who’s-who of business and industry in one place, but I think the aggressive marketing people and spammers sending everyone’s messages has driven the entire platform into background noise and a necessary evil for resume building and connecting with people in your physical community.
And you must remember LinkedIn doesn’t even feature in the top ten of global social media networks.