Using Analytics to Identify Affiliate Abuse – By David Naffziger

by FeedFront Staff on March 5, 2010

Detecting affiliate abuse is one of the more challenging aspects of affiliate program management.  Since the majority of affiliates add value to a program, identifying the abusive affiliates can feel a bit like finding a needle in a haystack.

The performance data provided by the affiliate networks is one of the most valuable aspects of the affiliate program, and can be used to identify affiliates engaged in abusive behavior.

Study the Inbound Referrers

Most visitors to your Web site come with information in their request that tells the Web site where the visitor came from. This information, known as the “referrer”, is used by programs such as Google Analytics to identify traffic sources.  There are three important patterns to look for:

  • Referrer from the Search Engines: If the referrer shows one of the search engines as the source, the affiliate is probably purchasing PPC ads and using an uncloaked affiliate link in their ad.
  • Referrer from Webmail Clients: Affiliates using email as a marketing tool will often show referrers from web email clients.
  • No referrer: A number of trademark bidders (or affiliates trying to mask the source of their traffic for other reasons) are using cloaked affiliate links with a referring technique that strips the referrer from the user’s session. 

Not all networks provide full referrers, however they often provide a report that at least shows the domain the traffic came from.

Understand Abnormal Conversion Rates

Affiliates using different promotional methods will have different conversion rates in your program.  Significant differences in conversion rate from your program average might indicate something is amiss. 

For example, unusually high conversion rates can be indicative of trademark bidding, while unusually low conversion rates (coupled with high traffic) can be indicative of adware.

Many merchants group their affiliates into categories (email, SEM, coupons, website, etc.), and then study the performance of the affiliates in each group.  This approach allows for a clearer identification of suspicious activity.

Study Transaction Timing

Just looking at when an affiliate’s transactions occurred can also be very informative.  Two things to look out for:

  • A temporary traffic surge: A quick, short burst of traffic could have come from an unauthorized email campaign or a successful social media campaign. Regardless, it is probably a good indication to look further.
  • Off-hours transactions: Transactions that only occur at night or on weekends might be indicative of a day-parted PPC campaign.

 There are plenty of good reasons why a non-abusive affiliate might be using a technique identified in this article.  Simply discovering an affiliate that matches one of the patterns above does not confirm the affiliate was doing something wrong. 

The best approach is to develop a relationship with the affiliate so that the merchant understands their promotional methods. These techniques should be used to identify affiliates that warrant closer investigation.

David Naffziger is CEO of BrandVerity, a firm that detects affiliate violations of merchant paid search policies.

Download the entire FeedFront issue 9 here – http://www.scribd.com/doc/24376105/FeedFront-Magazine-Issue-9
FeedFront issue 9 articles can be found here as well: http://feedfront.com/archives/article002334

{ 1 comment… read it below or add one }

Kevin April 28, 2010 at 9:48 am

I just posted a piece on identifying some affiliate traffic off of the Shareasale network using custom segments on my blog. Hopefully I’ll add the info for the other networks soon as I develop the methodologies….

Leave a Comment