Click Fraud in Paid Advertising, and What It Costs You | Simon Young and Steve Q

How much of your ad traffic is actually fraudulent?

Typically fifteen to twenty five percent of the traffic on Google and Facebook ads is fraudulent, bots clicking ads rather than real potential customers.

If you are running Google Ads or Facebook Ads, that is the scale of it. The main issue is that if you are optimising for a conversion event, a phone call, a form fill, a sale, the bots are clever enough now to optimise for that same conversion event. So your campaign, if it is not protected from fraud, is actually optimising itself to be defrauded, because you will get more conversions from fraudulent clicks than from actual potential customers. There are only really a couple of tools that help with this. I had been speaking to a company called CHEQ, spelled C-H-E-Q, based out of Israel, who as far as I knew were the only people worldwide stopping fraudulent clicks through Facebook.

Why would a bot bother clicking your ad in the first place?

Because there is money in it: fraudsters build large networks of sites, cover them in AdSense ads, and drive fake, clickbait traffic to them purely to collect the revenue.

I showed an example of an ad that had popped up in my Gmail, using a picture of Richard Branson and a claim that he had made someone five hundred pounds on television talking about bitcoin, sitting behind a web address dressed up to look like it belonged to the Daily Mirror. It is designed so that someone who does not know what they are doing sees a familiar face and a paper they trust and clicks through.

As a website owner you can put a display ad on your own site and get paid when someone clicks it, which is fine when it is legitimate. But imagine spinning up half a million sites with a bot, putting AdSense ads on all of them, and then sending clickbait traffic to them. The things that get clicked most are the obvious ones: certain kinds of images, dogs, and sensational made-up stories about public figures. Once someone lands on one of those sites, all they are there to do is click another ad, and the advertiser whose ad happens to be on that page is the one paying for a click that was never going to convert into anything. I was told about someone paying twenty five to forty dollars a click for a competitive commercial niche, office moving, and if fraudsters get ads like that onto half a million sites, that is a serious amount of money sat there generating itself.

Why is YouTube pulling budget away from Facebook for big brands?

Because it is still effectively free to reach people: you do not pay if someone skips within the first thirty seconds, and Google lets you target far more precisely than Facebook does.

We had been talking to national and global brands who have what they call a YouTube first strategy, because they can see people moving away from Facebook. One brand I cannot name was spending over a million pounds a month on Facebook and shifting everything to YouTube, because the targeting is so much better. On Facebook, if I want to target people in banking who are commercial customers rather than personal ones, I cannot easily do that. I have to overlay people who own businesses with people who follow banking pages, and it gets messy, because someone might have owned a business three years ago and no longer be relevant. On YouTube, I can target people who have visited a specific URL, say a bank's business section, and combine that with people who have searched for commercial finance in the last thirty days, and only hit people who have done both.

From there it becomes a funnel. People who skip the first ad are shown a different piece of creative or a different call to action. People who watch past a certain point, say two minutes, get shown a second video that goes further, and the ones who get furthest through that get the video with the actual sell in it. Four minutes seems to be the sweet spot for retention: after that, people start tuning out.

How do you stop the funnel wasting money on people who were never going to buy?

You do not eliminate them on day one, you eliminate them through evolving the campaign over time, because the qualification stage and the data from running several versions of a campaign tell you who is actually converting.

You would not get rid of a bad audience in month one, because you have to test first, and that is where a lot of people go wrong: they do not do the testing. Once campaigns are running, you can see that campaign A with one type of targeting produced a lot of time wasters, while campaign B, even though the cost per lead was slightly higher, produced better quality. I had an example running commercial finance leads where the client's typical cost per acquisition, to actually sign a piece of business, was over a thousand pounds. We tried a quick and dirty method on Facebook generating hundreds of leads at five or six pounds each, and a fully qualified funnel on YouTube taking people through the warm-up steps described above. On the YouTube funnel we were paying sixty or seventy pounds a lead, but converting one in three, which took the effective cost of acquisition from over a thousand pounds down to around one hundred and eighty pounds.

What actually works on LinkedIn at the moment?

A short, direct message straight to the top of an organisation, not a long automated sequence.

I still get twenty to forty automated inbox messages a day, and anything longer than one paragraph I do not even bother reading, and I am sure most people watching are the same. What works for me now is a very direct approach: a short message asking if they want some help or some insight, and if they do not reply, leaving it. That is getting me a fifteen to twenty percent reply rate from blue chip, C-suite level people. From there I move to a short first email asking for a bit of detail about what they want researched, and the ones who come back typically want to set up a meeting with their marketing department, which shortcuts the whole process. Generic automated openers that say things like "thanks for connecting, how are you finding these interesting times" get ignored, because it is obvious nobody wrote them personally.

How do you actually get a YouTube channel to be seen?

You have to pay to promote it, because YouTube will not organically surface a channel with no traction, and retention through a video is the strongest ranking signal you can get.

The internet is like a massive library: you can write a page, index it, and put it in the middle of a mass of books, and nobody will ever find it. I have put stuff on YouTube myself and, unless I have actually sent the link to people, it can sit there for three months and get one view. That is the catch: YouTube wants you to pay to promote the content, and paying is what gets you the organic traction afterwards. It also needs consistency, because putting up one piece of content and leaving it is the worst thing you can do.

There are roughly five to six hundred hours of footage uploaded to YouTube every single second we are alive, and most of it is not going to be shown ads. To break through, you need a paid strategy. I am doing work with someone who is a business celebrity, and part of her growth strategy involves five to ten thousand pounds a month on paid ads, which is leading to hundreds of thousands of subscribers. Once you have built that audience, everything you put out afterwards gets good traction straight away and it snowballs, but you have to go over the top of the hill first, and that initial investment is what a lot of people will not commit to.

The most important thing for ranking organically is getting people to watch as far through your video as possible, or to take specific steps within it. When we published this interview afterwards, we time-coded parts of it, so if someone only wants the click fraud section or the section on a particular topic, they can jump straight to it. Someone clicking through to a specific point and watching from there is as good a signal for ranking as you can get, and if you can then appear organically in the suggested videos alongside it, that is when a channel really takes off.