Attribution Fraud: Are You Getting Gamed?

sephzdarkoupdatedData-Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.

Today’s column is written by Seph Zdarko, director of data, modeling, partnerships and attribution initiatives at Quantcast.

The Oxford English Dictionary and Google tell us that the word “fraud” has Latin origin. Cheating people through criminal deception has a long history, going back millennia. But in the advertising industry, ad fraud – deliberately attempting to serve an ad that has no potential to be viewed by a human – is somewhat newer.

Ad fraud is real, representing about 10% to 20% of the digital ecosystem as a whole, although exactly how bad it is depends on where and how you buy. Suffice it to say that it is a serious problem for programmatic advertising, and the industry has started to take action.

Another type of programmatic fraud is particularly pernicious, but hasn’t received nearly enough attention. It’s called attribution fraud, and it refers to the deliberate practice of attempting to serve ads outside of an attribution model’s measurable capabilities.

For example, advertising platforms can game last-touch attribution by allocating the majority of the advertiser’s budget to retargeting consumers who are already close to a conversion. That greatly increases the likelihood that the platform will get the credit for the last touch on any subsequent conversion, but at the expense of delivering a lot of unnecessary retargeted ads, which come out of the advertiser’s budget.

Attribution fraud can affect 50% or more of marketers’ ad spend, depending on how and where they buy, and you probably have no idea that it’s happening. Like other kinds of ad fraud, it has been around for quite some time but it didn’t explode until real-time exchanges, big data and programmatic started to take off.

To understand why, you need to look at how attribution has evolved. Back in the early days of online advertising, buying targeted ads was a simple process. You bought sites directly and maybe made buys on a network or two. Simple, bottom-of-the-funnel measurement was all that was needed to determine basic ad effectiveness. Last-touch/click attribution worked well, and the use of more complex algorithmic attribution modeling was mostly for multichannel allocation. However, that was back when marketers bought ads 1,000 at a time across known inventory sources.

Today, precision targeting instruments can slice and dice individuals and impressions down to the millisecond. These new tools can dance circles around the outdated bottom-funnel attribution models. Unfortunately, it is often easier for sophisticated demand-side platforms (bidders) to game the bottom-funnel attribution models rather than work within them. Fraudsters can easily elude even the most complex regression and algorithmic modeling offered today.

Attribution fraud hasn’t become a hot topic like ad fraud is today simply because we can’t easily see it – yet. Attribution gaming is happening, but we’re not sure where and how much.

An open-source initiative called “split-funnel attribution” allows providers to augment their existing attribution models and share specific metrics about where fraud and manipulation are happening. Used by Ensighten, Abakus and other vendors, it adds an additional point of measurement into an existing attribution model at the point where the consumer first visits the advertiser’s site. That first site visit is also the handoff between the ad delivery platform’s third-party data and the advertiser’s first-party data.

Adding this delineation point allows attribution models to run separately on both the upper funnel (the prospecting phase, prior to the first site visit) and the lower funnel (retargeting phase, between the first site visit and the eventual conversion). It results in useful metrics such as split-funnel mix, which shows how much marketers are spending on retargeting vs. prospecting.

Splitting the funnel in this way enhances understanding of the causal relationships between the upper and lower funnel activities. It lets marketers figure out exactly how much each contributes to the eventual conversion event. It shows them the individual incremental gains from prospecting and retargeting separately and, at the same time, exposes the majority of attribution fraud.

Follow Quantcast (@Quantcast) and AdExchanger (@adexchanger) on Twitter.

Enjoying this content?

Sign up to be an AdExchanger Member today and get unlimited access to articles like this, plus proprietary data and research, conference discounts, on-demand access to event content, and more!

Join Today!


  1. Split funnel attribution seems to introduce the same issue where vendors can still game the last-touch model to receive credit for each conversion point. It doesn’t seem to change the issue, rather add a second conversion point where the issue can occur. Doesn’t it make more sense to look at first touch and last touch for each vendor in order to determine which strategies are driving incremental qualified consumers into the top of the funnel?

    • Thanks for the comment Greg. This is a common misunderstanding and while your logic would seem to make sense on the surface Split funnel is much more foundational than what you describe. Split Funnel doesn’t add a new conversion point. The first site visit is not a conversion event. It is just a transition point on the way to a conversion at the bottom of the funnel after the retargeting phase. You can still do first touch and last touch as you suggest, but now you can separate the two upper and lower phases in your reporting and modeling. I highly suggest you watch the video that is linked in the article above (only 18min) and you should get a better understanding of the how Split Funnel works and why it is in your best interest to implement it as a marketer. Only good things can come from greater transparency.

  2. Joe Ayyoub

    Great column Seph and I fully agree that as ad platforms are getting more sophisticated, it is getting much easier to game the traditional last-click or bottom funnel models. I am an believer in the Split-Funnel attribution initiative and think it can simply “fix” the bulk of this epidemic.

    • Thanks Joe. It is very unfortunate that some digital media companies feel they need to employ these gaming tactics to stay competitive. I can tell you first hand that over the last few years I have seen Split Funnel solutions implemented there is more often than not one or more partners on the plan that that are clearly doing things they aren’t supposed to be doing. My hope is that with more transparency and awareness we can start to eliminate attribution fraud, just like we are actively doing with ad fraud.

  3. Prickly Pete

    Showing a lot of retargeting impressions to establish last-view credit isn’t fraud. It’s rational behavior incentivized by last-view attribution, and entirely within the constraints and parameters. It is fair to say last-view attribution is flawed, but to call this behavior “fraud” is ridiculous.

    • Defrauded advertiser

      From the advertiser’s point of view, it is fraud. Any retargeting company that is “Showing a lot of retargeting impressions to establish last-view credit” but not to actually influence the target audience is defrauding the advertiser. The advertiser is paying for effective ads not gaming the system. Yes, from a strictly legal perspective it’s not fraud. Rational behavior by whom, the advertiser or the retargeting company? Which retargeting vendor do you work for?

    • Interesting how he talks about targeting low funnel site traffic as fraud. Honestly, it’s the first thing you should do. For example, if you’re working with an e-comm client, you would want to create remessaging programs around card abandonment first before working on site bounce. Why wouldn’t you apply the same logic to your consumer funnel?

    • Let’s not split hairs here, Pete. The Latin is fraus ‎(fraudis):
      1. A cheating, deceit, guile, fraud.
      2. A delusion, error.
      3. bad or ill intent

      I’d say the behavior the author describes fits any or all of these quite well!

    • Thanks for the comment Prickly Pete. Let me ask you this… Is deliberately bidding on bot traffic to increase campaign CTR considered fraud? Is cookie bombing from a toolbar or in a nested or non-viewable ad to game CPA considered fraud? If a media partner is hired to do prospecting only and they deliberately do retargeting to game their CPA without disclosing it, is that Fraud? Fraud has many levels of severity, but Cheating people through deliberate deception is the point. I’m not saying that last touch or other forms of bottom funnel attribution are fraud, I’m saying the bidders that are deliberately gaming them are. And last touch is the most easily gamed. If last touch is the only form of attribution a marketer is using, then they most likely have an attribution fraud problem. These bottom funnel measurement solutions can’t see the fraud and that’s what this article is about. My goal is to educate and encourage marketers to put another referee on the digital field that can spot this type of manipulative activity. It’s about transparency and removing the cheaters. With the increased transparency and insights offered by Split funnel everyone wins over the long run.

  4. Not sure split funnel actually solves the issue. Instead you just end up with two points where people can game the system. Wouldn’t a first touch vs last touch analysis make more sense? We’ve been using this model successfully for years to optimize the entire funnel, and measuring both on last touch and first touch (in relation to cost) doesn’t allow vendors to “game” the system.

    • Great question Greg. It is actually very common for marketers to initially mistake split funnel with a multi-touch model. Split funnel is much more foundational then any specific attribution model you use. All types of attribution can, and should, be applied to it. Last-touch, first-touch, multi-touch, regression, game theory or whatever. Split funnel is about delineating the upper funnel (prospecting) from the lower funnel (retargeting) at the 1st and 3rd party cookie hand off during the first site visit. If you watch the video that is linked in the article above (only 18min) you should get a better understanding of the nuances. Companies like Abakus for example split the funnel and then run game theory across both upper and lower independently. Companies like Ensighten or Quantcast run a multi-touch or last touch on each. Whatever attribution model you use the point is that you split the data at the handoff point where the vast majority of the attribution fraud happens. If you just run your attribution model across the entire funnel, both upper and lower together, not independently, you run the risk of not seeing the fraud and getting gamed unknowingly. Will implementing a split funnel solution solve 100% of the potential fraud? No, just like ad fraud nothing will ever reach 100%, but split funnel will shine a light in the areas where the majority of the fraud happens and likely knock it down to the low single digit percentages. I’m happy to geek out on the topic with anyone that is interested. Please feel free to connect with me and start a conversation.

    • Close to a conversion means after the first site visit at a minimum, but it generally means after several visits or other on-site browsing indicators. An online conversion cannot happen until after the first site visit. It might happen on the first site visit, but typically there are more than one site visit and a fair amount of retargeting that happens along the way towards the conversion. As consumers visit the site more often and engage more with products (potentially even putting items in their cart) they get closer to a conversion. Most bidding platforms increase the frequency and price for ad opportunities as consumers get closer to the conversion event to increase the odds that they get the last touch credit. A little bit of retargeting is needed here to keep the conversion front of mind with subtle reminders to the consumer, but without proper management this is where unwatched retargeting can get a bit crazy. With a bottom funnel measurement solution all the media partners on the plan often fight to steal the last touch (or click) credit. This can result in hundreds and even thousands of ads being served to a single consumer in very short period of time. Too many ads being served to anyone is a bad thing for everyone involved. The marketer, the vendor and the consumer all lose in this scenario. With even a simple last touch Split Funnel insights report a marketer can keep tabs on who is doing what and keep all partners on a media plan in check, for the marketers benefit and the consumers.