Click Fraud: Using Attribution to Mitigate Risk

Thanks to the influx of the Internet over the past decade, we have all have heard the nightmarish stories of online gone bad, from tales of major security breaches at top retailers to website hacks, spammers and related blunders. And fortunately, marketing departments have gone largely unscathed — until now.

 

Click fraud is a growing internet crime costing marketers across the globe millions in lost spend. In fact, Alex Mindlin wrote in The New York Times that 25.8% of fraudulent ad clicks are in the United States and 44.1% of ad clicks originating from Vietnam are fraudulent.

 

Click fraud occurs when a person, automated script or computer program imitates a legitimate user’s click on an ad, for the purpose of generating a charge per click without having actual interest in that ad. Sometimes it’s done by fraudsters who want to steal a portion of the advertising budget of the marketer.

 

They either manually click on the links from several computers, use a computer program to imitate the manual click and deploy it on several computers, or worst of all, use malicious programs to spread these imitating scripts across several computer networks and use Trojan code to turn the average machines into zombie computers that would run the scripts to generate revenue for the scammer.

 

In some cases, it’s not fraudsters, but a known competitor. It’s sneaky and immoral, but these competitors see no problem in clicking away at ads in order to quickly deplete daily budgets of others in the market, opening the door for them to bid low prices at the prime time, with less competitive pressure.

 

Most marketers know that click fraud exists, yet very few measure it and fully understand how it affects their campaigns. It is high time for marketers to understand how much is at stake.

 

Know If — and How — You’re Affected

 

In analyzing Visual IQ’s customer base, we have learned that marketers are losing about an average of 16.7% of their PPC budgets to the fraudsters every day. Some marketers even lose up to 45% of their budgets without even knowing about it.

 

Marketers can figure out if they’re being affected by looking at the performance of campaigns. Low conversion rates, trivial ROI on PPC campaigns, and lagging behind the competition are signs of click fraud.

 

The percentage of click fraud to the campaign budget varies from marketer to marketer. Each marketer should make their own assessment on how much the click fraud affects their business.

 

Take Simple Steps Towards Prevention

 

Prevention is the best method to deal with click fraud. It can be pretty difficult to prove click fraud and it is more difficult to get back the money lost. Instead of looking back with regret, be proactive about preventing click fraud using the following simple methods:

 

• Make sure that your PPC campaigns are limited to the geographies where you sell you products and services.
• Avoid PPC campaigns in the geographical areas that are prone to more click fraud incidents, like Vietnam and Nigeria
• Have daily budgets. See how you spend your daily budget by hour by hour and flag any activity that looks suspicious.
• Compare the conversion ratios of your PPC campaign with other campaigns.
• Tune campaign parameters on a regular basis to avoid potential click fraud.
• Tackle Click Fraud With Attribution Technology

 

While the above methods are a good first start, the best way to prevent click fraud is by tracking your clicks and analyzing their patterns. Many of today’s attribution technologies can implement click fraud measurement and analysis.

 

These technologies remove a lot of the manual work above, while backing up potential fraud with scientific proof. The most effective technologies leverage attribution to prevent click fraud by conducting the following steps:

 

1. Collecting detailed attributes of every click, including the keywords, geographic locations, IP addresses, time-of-the-day, domains, ISP and publishers.

 

2. Collecting the engagement stack of each of user. Engagement stack of individual users includes the timestamps of every impression, click, conversion and consumption of ads from different channels such as email, online display, search and affiliates.

 

3. Analyzing the click attributes and engagement stacks to find the clusters of potentially fraudulent clicks.

 

4. Grouping these clicks into categories like high/medium/low propensity to be fraudulent.

 

5. Quantifying the damages made by each category — including assessment of damages to the business and losses to the campaigns.

 

6. Using advanced modeling techniques to find the patterns of publishers, geographical areas, IP address groups and ISPs that produce more fraudulent clicks.

 

7. Based on the analytical insights from the models, generating recommendations to the media planners to prevent fraudulent clicks. This step is vital for marketers to effectively take action to mitigate potential violations

 

While it is impossible to prevent 100% of the click fraud, the problem can be highly contained. Marketers with advanced attribution technologies are preventing up to 70% of it and saving millions of dollars every year.