Click Fraud: The Peril of Pay per Click
As a follow on to our series on pay per click advertising, one question that I’m asked a lot about is about click fraud – how do you know if your competitors are clicking on your ads and what can you do about it? Since it’s quite a complex subject, we’re linking to several articles on the subject as well as showing the advice from the Google Adwords Help Centre, which gives in depth explanations about the problem.
So what is click fraud?
Click fraud is one of the perils of pay per click advertising. Basically it involves third parties clicking on your ads multiple times to use up your budget quickly so that your ads are stopped also making your business suffer so that your paid ads deliver no return from your spend.
Fraudulent clicks can come from a variety of sources: competitors, disgruntled customers, disgruntled suppliers, even disgruntled staff all trying to increase your advertising bill as much as they can, to the point that you may need to stop your ads.
The major pay per click advertising providers, i.e. Google Adwords, Yahoo Search Marketing, Microsoft AdCenter are all well aware of the problem and there has been a lot of publicity about the subject over the last few years.
Most people have 2 burning questions on the subject:
What are the search engines doing about it?
What can I do about it?
Firstly, what are the search engines doing about it?
The providers have systems in place to track fraudulent clicks or ‘invalid clicks’. They can track multiple clicks from the same IP address, which is often how click fraud is detected. The pay per click provider is then obliged to refund those clicks back to the advertiser. This is Google’s response to ‘invalid clicks’ directly taken from their Adwords help section:
“How does Google detect invalid clicks?
The security of Google AdWords advertisers is important to Google and we have dedicated a number of resources to protect your account against invalid activity:
Detection and filtering techniques: Each click on an AdWords ad is examined by our system. Google looks at numerous data points for each click, including the IP address, the time of the click, any duplicate clicks and various other click patterns. Our system then analyses these factors to try to isolate and filter out potentially invalid clicks before they ever reach your account reports. This detection and filtering occurs over a number of levels including the following:
- Real-time systems filter out activity fitting a profile of invalid behaviour (such as excessively repetitive clicks)
- Clicks and impressions from known sources of invalid activity are automatically discarded
Advanced monitoring techniques: Various unique and innovative methods are applied at each stage of the filtering process, thereby maximising proactive detection of invalid activity. Our engineers are also constantly improving our monitoring technology, enhancing filters and examining a growing set of signals.
The Google Team: In addition to our automated click protection techniques, we have a team that uses specialised tools and techniques to examine individual instances of invalid clicks. When our system detects potentially invalid clicks, a member of this team examines the affected account to glean important data about the source of the potentially invalid clicks.
One of the goals of our team is to make invalid activity very difficult and unrewarding for unethical users, thereby decreasing their chance of success. Additionally, if we find that invalid clicks have been charged in the past two months, we will credit advertisers' accounts.
Dr. Alexander Tuzhilin, an independent expert who has examined our detection methods, policies, practices and procedures, has documented these and other details of our monitoring system. You can find his report here: http://googleblog.blogspot.com/pdf/Tuzhilin_Report.pdf (pages 25-26, 47). (English only)"
What can you do about it?
The best way is to monitor your pay per click campaigns and look out for any sudden spikes in traffic or unexplained trends in your clicks that cannot be accounted for. For example, sudden very high click through rates, particularly on your more expensive competitive keywords which usually have lower click through rates. Perhaps you usually receive a high ROI (return on investment) from your adwords campaign, but suddenly the return is dimished. If you think that you are the victim of click fraud, the first thing you should do is collect your data and contact the pay per click provider with your findings.
Google has the following advice on what you can do (taken directly from the Adwords Help Section):
"What can I do to help monitor or prevent invalid clicks on my ad?
Google's proprietary technology is designed to filter out invalid clicks and impressions before they ever reach your reports. However, here are some things you can do yourself to help identify and monitor suspicious activity:
Check if your ad is receiving two or three times the normal number of clicks you would receive on a single day. Keep in mind that you may see a sudden increase in impressions or clicks as a result of legitimate user activity, such as seasonal increases in clicks or the appearance of your ad on the Google Network.
Look at click patterns over time. Identify changes from week to week that are not explained by any of the legitimate reasons for high click volume. Run an Invalid Clicks Report to view numerical data on any invalid clicks already filtered out by our system. You do not need to do anything about the invalid clicks data reported and you will not be charged for these filtered clicks.
Compare your web logs to your AdWords account reports. (If you are not sure how to access your web logs, talk with your webmaster or website hosting service.) Your site's web logs will often contain information on the IP addresses of the users that are clicking on your ads. While this can help give you an idea of who is clicking on your ad, remember that certain Internet Service Providers (ISPs), such as AOL, assign the same IP addresses to a large number of users. As a result, multiple visitors may access your site from identical IP addresses, giving the impression of repeated clicks from a single visitor. Learn more about using your web logs to monitor click patterns.
Use tracking URLs and referrer headers to monitor and evaluate your web log reports. This can help you obtain more relevant data from your web logs. Learn more about these tools below. (Web servers vary in how they format their logs, so the following information may not apply for all advertisers.)
Tracking URLs
Tracking URLs can help you identify all the traffic that comes from Google (or other sources) to your website. In many web server logs, a page visit entry would be identical whether a user clicked your AdWords ad on a Google search result page or your website's Google free search result listing. In order to distinguish which clicks come from your AdWords ads, you can give your ad a tracking URL.
To take advantage of tracking URLs, place a parameter at the end of your Destination URL. For example, if your URL is www.yourdomain.co.uk, your tracking URL could be www.yourdomain.co.uk/?source=googleadwords.
It is important to test each new tracking URL in your own web browser to verify that it is linking properly to its specified page. If you find that a tracking URL is not linking properly, you might want to eliminate the forward slash after the domain. For example, instead of www.yourdomain.co.uk/?referrer=googleadwords, try using www.yourdomain.co.uk?source=googleadwords.
Once you have created your tracking URLs, look in your web server logs or your third party tracking software to get traffic data for your ads. If you are using your web server logs, look for the tracking URL in the GET entry to identify which website visits came from your Google AdWords ads.
Referrer Headers
Referrer headers show the last page the internet browser loaded before loading your website; this typically indicates the site that users were on when they clicked on your ad. For example, if a user searches on your keyword on Google.co.uk and then clicks on your AdWords ad, the referrer header would most likely read 'http://www.google.co.uk/...' Conversely, if the click-through had come while your ad was appearing on AOL, the referrer header would read 'http://www.aol.com...'
While it will not tell you specifically who is clicking on your ads, Google Analytics can help you learn more about user behaviour on your site. You can use Analytics reports to determine which sites users are on when they click your ads, what they do once getting to your site, which cities such users are located in and more. Learn more about Google Analytics.
If you believe that your account reports reflect click activity that is more extreme than ordinary user behaviour or that exhibits strange patterns, please contact our click quality team."
Although we’re showing Google Adwords answers to the problem, the other providers have similar advice to advertisers.
We've found several articles on click fraud which make interesting further reading on the subject along with findings from campaigns that have been hit by click fraud:
How Click Fraud Could Swallow the Internet - from Wired.com
Click Fraud: What It Is, How to Fight It - from Clickz.com
Exclusive: Google’s Click Fraud Rate is Less than 2% - from MarketingPilgrim.com
Click fraud scams on the increase - from ZDNet.co.uk
Needed: A Detailed Picture of Click Quality - from Traffick.com
Click Fraud Detection - from MarketingExperiments.com
|
