Identifying Click Fraud Patterns

In Part V, "Tracking Your Return on Investment," I discussed how to track the various components of your paid listing campaigns: search engine, keyword, ad listing, and landing page. Assigning a unique tracking code to every listing helps you determine the ROI of each ad component. Furthermore, this enables you to spot unusual fluctuations in your campaigns that could indicate click fraud activity. Although you can identify suspicious click fraud behavior by relying on raw log activity shown in your web analytics reports , you cannot effectively attribute such activity to "questionable" clicks without having tracking URLs in place.

Look for anomalies per search engine and per keyword by studying the following data:

• Average daily clicks

• Average page views per click

• Average conversion rate per keyword, per click

If possible, monitor:

• Click trends by hour

• Click trends by day

• Click averages based on specific positions

Abnormal behavior could include:

• Spikes in click volume for specific terms

• Click increases plus atypical visitor behavior (zero conversions, zero page views, and off-peak click times or dates)

• More than one competitor dropping out of a top position during times of suspicious activity, while one competitor remains seemingly untouched (remember the story of Bob, the Evil Clicker)

These are the patterns Jessie Stricchiola, now president of the search engine marketing firm Alchemist Media, Inc., monitored while she managed a large paid listing campaign for a nationwide law firm. She had worked closely with the firm's in-house PHP programmer to develop a tracking system that would track all click data, independent of cookie and IP address (known as session tracking ). Each keyword had its own unique tracking URL that was given to the search engines that stored click data, including keyword and the specific engine source in the database. She studied the click patterns and in one instance was able to associate click fraud activity with specific IP addresses. Between FindWhat.com and Overture, Stricchiola tracked nearly $10,000 of click fraud within 30 days.

Important

An increase in advertising efforts or media attention could contribute to traffic surges from paid listings. And web site modifications can influence visitor behavior. Evaluate other possible factors before you assume there's click fraud activity.

Outsource your ROI tracking efforts if setting up a home-grown system seems too time-consuming . A few third-party tracking vendors even incorporate a form of click monitoring. But you might want a program that's dedicated for this type of reporting. Thankfully, a click fraud service made its debut in the summer of 2003, and it seems promising .

WhosClickingWho? is an independent pay-per-click (PPC) auditing service designed to gather information on all visitor clicks to a web site via all major PPC search engines: Overture, FindWhat.com, Sprinks, 7Search, ah-ha.com, ePilot, GoClick, LookSmart, Google AdWords, and all of their affiliates . WhosClickingWho? compiles and generates the following data:

• Who is hitting your site from multiple pay-per-click search engines.

• Uniquely identify each PPC visitor via IP address and unique identifier string, enabling you to detect most AOL, Prodigy, and even dial-up abusers who have multiple IPs.

• Generate custom reports detailing who has clicked on your site from PPCs two, three, four, five, or even six times, allowing you to identify potential PPC abusers quickly.

• Download your reports or email them with comments.

Figure 21.1 shows how repeat visitors are identified by their IP addresses through WhosClickingWho?.

Figure 21.1. WhosClickingWho? is a pay-per-click auditing service designed to track and report click fraud activity.