AdWords
6.7K members online now
6.7K members online now
Improve your AdWords performance and boost your ROI, CTR, and Quality Score
Guide Me
star_border
Reply

Let's talk about click fraud

Follower ✭ ☆ ☆
# 1
Follower ✭ ☆ ☆

I'm currently experiencing what I believe to be a very sophisticated case of click fraud. 


Click fraud is one of the most under-discussed subjects in the AdWords community. The majority of discussions on the subject, even by the authority blogs appear to stem around the most basic kind; what to do when a competitor idly clicks your ads, which isn't really an issue anyway. Googles own prevention does a good enough job at nipping this in the bud.


Though I had never experienced click fraud personally, at least not knowingly, I always wondered how often the scales would tip in favour of the fraudsters. Moments in time when their technology supersedes Googles defences. Much like hackers vs Norton and Mcafee.


I have been working in this specific niche for just over two years and have achieved very good, consistent results in a particularly aggressive niche. My competitors are Fortune 500 companies and so it goes without saying the reward for taking competitors out is HUGE.


The account is in good standing; 8-10 QS on all keywords, consistent 25-30% conversion rates. CPC, CPA, impression share etc are all healthy. 


The past month conversion rates have been suddenly tanking to ~10%. CR has dropped as low as 3% on some days, something I didn't think I'd ever see, not without a major on-site issue.


I have checked all the usual; tracking code, page speed, account history, landing pages, traffic volume, competitor offers, ad copy, increase in device specific traffic etc etc etc... I exhausted the possibilities before considering the possibility of foul play.


Analysing the data from AdWords, analytics and LuckyOrange, the main change on the days where the conversion rates tank, is the time on site would also drop from avg. 1:45 to 13s for example. One could argue that this is a natural phenomenon from the way in which Analytics measures session duration, but Lucky Orange data supports the big drop in session time.


Other noteworthy observations; increase in CTR and the budget would also be eaten up faster than usual. 


Because the keyword > ad copy > landing page is laser focused, a typical user would at the very least explore the landing page offer, with the majority at least filling in the first stage of the lead-gen form. However on the days where CR tanks, Lucky Orange shows the majority of sessions remain static for 5-15 seconds and the session ends. Some of the sessions may see the odd scroll or mouse wiggle, but clearly no interest in the content. Entirely different from any typical day from the two years previous.


I have whitnessed several sessions whereby the landing page URL would switch from one landing page to another in real-time. The landing pages are not linked to one another. Had the user clicked back and performed a new search, a new session would have been recorded. I'm not entirely sure what this suggests, but it certainly raised an eyebrow.

 


I contacted an online security company who said my data heavily suggests click fraud. They claim that softwares exist that spoofs IP, randomises device, OS and browser, in addition to attempting to mimic human behaviour. Essentially sophisticated bots. They also claimed that there are also "click farms" where humans, rather than bots are paid by an advertiser to click on competitor ads with the aim of taking them out of the auction.


The company in question offer quite an expensive solution that uncovers bots trying to mimic humans and compiles reports to present to Google as proof of click fraud. The software is much more elaborate than the likes of Clickcease for example, which frankly is a waste of time/money.


However, the very fact the company offers a "solution" one could argue that they're bias towards the idea of the existence of bots and click farms sophisticated enough to go undetected by Google.


In an attempt to find out whether we are being targeted specifically or specific locations in the auction I added a script that targets position 3.5. It has only been 24 hours but conversion rates have not only returned to their former glory, but beyond. However, the data is currently too thin to make any conclusions. Tomorrow could be another story.


I don't imagine many companies would have been able to survive the onslaught that we've faced this past month. It has exhausted hundreds of hours of time and hit us hard financially.


Fwiw I reported the issue to Google, who were quick to inform me that no click fraud was found, followed by a canned response about their elaborate measures to protect advertisers. I appealed and this time they suddenly found 60 invalid clicks, though a positive result, only a fraction of what I believe to be invalid.


I gathered more data and more evidence, to be told that policy states an advertiser can only submit one click fraud investigation per month.


Googles ignorance to the possibility that such software is being successfully used and going unseen by their defence mechanisms, only makes them and us more vulnerable. They should be working closely with advertisers to ensure they stay one step ahead of the game.


Do any of you have similar experiences? What did you do to combat the problem? How did you get on with reporting the activity to Google?


I appreciate there's going to be a decent percentage of you who will think it's blasphemy to dare suggest that click fraud exists.


Many thanks for reading my lengthy post. I look forward to any insight you might share.

1 Expert replyverified_user

Let's talk about click fraud

Rising Star
# 2
Rising Star

Hi Paul,

 

You are not the first to bring up this topic here. Your analysis of the situation seems to be correct. Unfortunately there is no solution that we or Google can provide to stop it happening. Google does listen to such complaints and can provide refunds if sufficient evidence is provided.

 

I've witnessed this many times in my career and it tends to occur to two groups of people - tradesmen who take personal affront to someone else having the #1 ad, and industries where people are making "easy money".

 

In less sophisticated cases, sometimes avoiding certain geographical locations, or devices, or times of day can work. Or just pausing your ads for a week or two until they forget about you. 

 

 

 

Let's talk about click fraud

Follower ✭ ☆ ☆
# 3
Follower ✭ ☆ ☆

Hi Rob, incidentally this is a trade niche. Our competitors are British Gas and the likes. 

 

I am hoping that the software ware provided by the company I mentioned above can give me sufficient evidence to at the very least receive a refund on the fraudulent activity. 

 

It is frustrating when speaking with AdWords advisors who try to explain the activity with logic such as seasonality or day of the week. Conversion rates don't just tank overnight to this degree. Plus we have years of data and know exactly when to expect dips, but sudden drops from 25% to 3% CR don't add up. 

 

Taking ourselves out of the limelight by bidding less aggressively is showing early promise. But I am not particularly optimistic after the performance issues over the past month and the sleepless nights that ensued. 

Let's talk about click fraud

Rising Star
# 4
Rising Star

So I'm guessing that you or your competitors basically just collect leads and sell them on? If so, there could be competitors who also use other channels to get leads, some of them not white hat, and they have an expectation of "easy money".

 

I'm just trying to point out that, in my experience, this behaviour is quite rare, and not normal competitor behaviour. 

 

 

Let's talk about click fraud

Follower ✭ ☆ ☆
# 5
Follower ✭ ☆ ☆

Hi Rob, no in this instance we are generating leads for our own purpose. Though I do have future plans to sell leads on. 

Let's talk about click fraud

Follower ✭ ☆ ☆
# 6
Follower ✭ ☆ ☆

Really surprising that Google is missing even blatant click fraud.

 

Today one of the keywords was positioned 2.2 because the min-bid I'd set in the script was higher than the CPC for pos 3. 

 

I woke up to 5 clicks on this keyword, with a whopping 35.71% CTR. Each session of course lasting seconds. 

 

Exhausting.