Finally, your point about not sensationalizing the fraud point - that seems a little disingenuous and I have to call you out on that one. In your press release (point 1 - you put out a press release) you say that these are “initial test campaigns” but nowhere do you provide any countervailing stats to indicate that these figures are not representative. Your subhead is “Initial test campaigns using AdXpose demonstrate more than half of the impressions delivered and 95 percent of clicks came from suspected fraudulent sources” and the quote is “Click fraud and impression fraud is far more pervasive than the industry has been willing to admit”.

I agree with you that daisy-chaining is a huge issue and needs to be addressed specifically. And I think that many people reading the release know that it benefits you guys (which doesn't make your argument flawed BTW - one of the biggest logical fallacies we fall prey to is discounting someone's arguments because their point is in their interest) and discounts it vs. an industry analyst type of report, but clearly your hope is this gets picked up and there are lots of hooks for it to do so and sensation doesn't hurt.

We've been working with folks in the industry on initiatives around this as well, and would love to collaborate with you guys and others to more effectively create some best practices in the industry. It will benefit our advertisers for whom we work every day.

Thanks for the thoughtful review of and response to the study. I would expect no less.

I did want to respond to some of your concerns. First, this statement: “Certainly, you shouldn’t create an experiment and call it representative of the marketplace when your experiment by design will seek out the very thing you are looking to measure.”

It should be boted that this was not an “experiment”. This was simply the data we collected during the period in our normal course of business of buying media for clients.

Secondly, the goal was not to paint the entire marketplace with this brush. The fact that some news outlets read the report and saw fit to sensationalize it is (while an admittedly happy PR development for us) is beyond our control. If you read the methodology (as Leigh Morrison and Kevin Lee assumably did before publishing their stories) you would see the following:

“Mpire conducted 11 RON buys across nine different ad networks (directly and via one exchange) in July. The buys delivered more than 20 million impressions, to ads from 53 different advertisers. These impressions were filled by the initial nine ad networks via downstream daisy-chaining on at least 45 additional ad networks on more than
100,000 sites.”

As you note in your post, this is a small and skewed sample set. We made this fact pretty darn transparent in the report. Anyone who takes this sample set and assumes we are claiming to represent the entire ad network universe is missing the point. We conducted buys during our normal course of business, wrapped the ads with our analytics code, and shared the results. You can and should interpret the data with your own filters based on your own experience – this is a data point, not a universal declaration of fact.

Maybe some clarification will help here: We believe the real issue is with “daisy-chaining” – that’s where we collected most of the data and hence that’s where we believe the highest volume of fraud occurs. I know for a fact there are top 20 ad networks who fill their buys via exchanges, and despite their best efforts, fraud and bad content issues still occur. Why? Because fraud and obfuscation are endemic to daisy chaining and create huge blind spots for bad actors to exploit. Therefore, we believe that daisy-chaining is irresponsible and does in fact make those networks who do it complicit in waste and underperformance if not outright fraud.

On the CPM versus CPA versus CPC issue: you are right to assume the lion's share of the results came from CPC buys. I am not sure how this invalidates the conclusions, however, especially given that poor traffic/fraud was prevalent on the CPM buys as well. While you may not be surprised by the learnings here, we feel a majority of advertisers will be, as witnessed by the strong response to and demand for the research and the product.

Further, your tone implies that one should expect to be defrauded if one buys CPC or CPA inventory via exchanges - that those who do purchase CPC are getting what they deserve. That's news to us, and we're sure to a lot of other, less sophisticated buyers.

All in, however, I think you make some valid points. We aren't a research shop nor are we former analysts, and it may show. But the data should still speak for itself.

Thanks again for your patronage and your candor, Rob. We need dialog around these issues more than ever, and anything that helps shed light and bring the topic of quality and safety for advertisers and publishers to the forefront is a good thing.

Kirby Winfield
CRO
Mpire