Bad quality instead of fraud


fraud_small.jpgADOTAS EXCLUSIVE — Does a trend exist in the online advertising industry for advertisers to label bad traffic as fraudulent traffic?

Advertisers may be conditioned to scream “Fraud!” instead of “Bad Quality!” like the public was trained to yell “Fire!” instead of “Help!” when attempting to attract attention.  Are advertisers crying wolf when it comes to fraud? Or, is the fraud simply bad traffic?

An advertiser’s fraud claim demands attention because chances are a liability has been created for the advertiser or the partner, or both. Whereas, an advertiser’s claim that the traffic is not backing out for them may command less of their partner’s attention. The former is an elephant and the latter is an ant.

As an advertiser,  I have experienced first-hand a lack of partner response when I have communicated quality concerns, in my case new member quality for our loyalty program Here’s how those conversations would unfold:

Me: Our click to conversion rate is quite low.
Partner: Maybe your landing page needs to be optimized.
Me: The return on investment and activity level of these new members is low.
Partner: It is only my responsibility to provide you the member; it’s up to you to monetize them.

At the time, had I known that using the “F” bomb would have been a powerful enough call to action, I may have just dropped it. In this example, the traffic I received was bad, but perhaps the new members I received came from a sub-publisher three levels away and the pattern of bad traffic was actually an indication of fraud. Could I have made changes and tested landing pages to improve conversions? Yes. Could I have spent more time monetizing the leads I was receiving? Yes. Could the traffic have been fraudulent? Yes. All I really wanted from my partner was a commitment to investigate, and I did not receive that.

On the flip side, the experience my organization has as an advertising network working with advertisers has provided me with a glimpse into a bigger issue. We do not share a common definition of fraud. The following is a snapshot of new clients that were on-boarded last week and the definition of fraud each of these clients provided:

Monday’s Fraud Definition: Duplicates
Tuesday’s Fraud Definition: Invalid credit card
Wednesday’s Fraud Definition: Cancellations on orders
Thursday’s Fraud Definition: Returns
Friday’s Fraud Definition: Numbers in names, etc.

It is important that we come together as an industry and collaborate on the real definition of fraud as it relates to what we do and the challenges we face. Each lead, sale, or transaction can be placed into one of three categories: Valid, Invalid or Fraud. Too often the invalid leads are being placed in the fraud category. The way to prevent and detect fraud is to first understand what fraud is.


  1. I like that your post acknowledges that this industry is founded on multiple partnerships and honest communication. The openness to state your needs (whether it be as an advertiser, as a network, or as a publisher) is the basic glue that holds this industry together. Don’t only focus on temporary differences in needs. Be open to discovering long-term, mutually-satisfying solutions.

  2. My company, Anchor Intelligence, works with ad networks, search engines, advertisers, and agencies to improve the quality of their traffic. By monitoring traffic quality and click fraud across its portfolio of clients, Anchor has developed a unique perspective on this issue.

    We agree with you on the importance of developing a standard, industry-supported definition of click fraud. Currently, Anchor Intelligence defines click fraud as clicks (or impressions) originating from the malicious intent of the user that have no economic value to the advertiser. As such, double clicks, traffic from known crawlers, and internal test clicks do not qualify as fraudulent clicks and should not be identified as such. Anchor categorizes both fraudulent clicks and clicks that are neither fraudulent nor legitimate as non-billable, but differentiate them by labeling the former as malicious and the latter as non-malicious.

    We believe advertisers should not be charged for fraudulent clicks or invalid clicks – hence, the “non-billable” label. However, we choose to distinguish between the two to enable customers to reduce their exposure to click fraud in the future. For instance, if an ad network knows which clicks are fraudulent and from which sites those clicks are coming, it can find and evict publisher sites that are perpetrating click fraud. Correct categorization of clicks saves time and resources and ultimately improves advertiser ROI, making it an imperative endeavor in today’s economic climate.

  3. Great article!

    To Toby’s point, I think the bigger issue in the direct response world is the persistent issue of lead quality.

    The question of “fraud” can be boiled down to two categories: 1) What the marketer did, and 2) What the consumer did. In both cases, it is the intentional action on the part of the individual that translates into the quality of the lead.

    In the first category, you may have an affiliate marketer doing something like distorting the copy or targeting suppression. As such, the resulting lead can be argued to be illegitimate, as well as illegal.

    In the second category, you may have a consumer trying to take advantage of incentive programs and submitting bogus information. This lead too can be argued to also be illegitimate.

    While the first category can be more clearly deemed fraud, the second one involving consumer action is more murky. Even if the advertiser cries “Fraud!,” there is still the lingering issue of whether or not the marketer has done his honest job.

    Should the marketer get his commission? Is this really “fraud” or just part of doing business online?

    It’s these kinds of gray situations that spark the most F bombs among networks, clients and marketers. I agree, until there is an agreed-upon ontology, lead quality and culpability issues will not be able to be addressed systematically.

  4. I posted a comment here on 2/27/09 and it hasn’t shown up yet. I wrote to Adotas the next day and never heard back. I came back today and the comment is still not here. So I tried to repost the comment, and it rejected the comment as a duplicate! But it’s not even posted here? So what gives?

  5. Chuck,
    I have no idea. Your comment here posted, and I’m the one who would get any email about missing comments. I don’t remember it, sorry. I also checked through our spam filter and couldn’t find it. If you have any more problems, you can reach me at editor

Leave a Reply to valencio Cancel reply

Please enter your comment!
Please enter your name here