The Importance of fraud verification - What publishers should know
Updated: Jun 18, 2020
In 2017, Uber filed a lawsuit against Fetch Media for allegedly buying fake clicks and taking credit for app installs it didn’t deserve. Fast forward to 2019 and Uber moved up a gear by going after the ad networks used by Fetch to place Uber’s media budget over the same time frame. It was, and still is, a landmark case, and one of the very first examples of a high-profile brand suing its media and ad tech partners over ad fraud.
Uber is not alone. Far from it. Juniper Research forecasts that digital ad spend will hit $520 billion by 2023, but last year alone the fraudulent component of that spend totaled $42 billion. However, managing fraud verification and detection products isn’t easy. Most companies in the media industry simply don’t have the resources to deal with it. But there are steps companies can take - whatever their size - that can minimize their risks.
Below I will provide an overview of how fraud verification companies work and what the real role of accreditation organizations is. My aim is to highlight the current state of the online ad industry and outline why publishers have become the side that suffer the most. My second goal is to describe the different tools and approaches available to minimize the risks.
Unbalanced Relations Between Publishers and Advertisers
There is a fundamental problem with this relationship. The media industry has created a system where publishers provide the inventory but, despite this, have only limited influence. Publishers are also partly culpable - they are not proactive partners against fraudulent activity led by the demand-side players. The communication between them is limited.
Lack of knowledge or delayed feedback leaves publishers in a vulnerable position. There is no official guideline to deal with these issues, which leads advertisers to choose a strategy of "Race To The Bottom"* on the cost of audience traffic. *After fraudulent activity, an advertiser will demand that a publisher lowers the cost of its traffic rather than cooperating to clean and resolve the traffic sources (Google benefits the most from this situation as they are considered to be ‘reliable’ and ‘qualified’ partners).
The fight against fraudulent activity puts a lot of pressure on publishers. Demand players often try to avoid the proactive blocking of fraudulent traffic sources before any official feedback is given. Once results are in (usually months after an event), only then will publishers be able to block the traffic.
Moreover, publishers will not receive any new solutions, tools or advice from the demand side. Instead, publishers will take action against specific sources, based on specific feedback instead of implementing a general solution. This approach creates anomalies in the results. For example, a block of IP addresses will cause a drop of specific traffic which does not necessarily eliminate the fraud source.
When publishers use verification products, they should verify that the product in question is data science based (a process in which the verification company collects data and updates internal policies and algorithms accordingly). It means that traffic sources can be highlighted as negative sources on an ongoing basis. Publishers who work with flexible products that are changed frequently can have a much better chance of reducing ad fraud whilst also lowering the risks of false blocking.
Don’t forget that professional fraudsters adapt their capabilities faster than most verification companies develop their detection systems. This won’t change anytime soon, and publishers who want to be on the front foot need to recognize this.
Lack of Regulation
As of 2020, there are no official rules or guidelines for digital ads and their users. Or put it this way, any guidelines that do exist don’t offer any protection. It’s somewhat baffling that there are many rules and laws covering users’ privacy, but nothing about handling fraudulent activity. This structural void is taken advantage of by industry giants like Google and others - after all, as there is no effective regulator, whoever has more resources sets the tone.
Today, many leading brands put great emphasis on protecting their reputation (as they should) but they are often too aggressive, heaping even more stress on publishers. The fear of rebuying the same traffic after a fraud issue is detected paralyzes buyers. Their actions are based on fear and the resulting wrong assumptions often lead to irrational decisions.
The Role of Verification Companies
A publisher’s main focus should be about ‘necessity’. What is the reason for verification companies to collaborate with publishers? After all, these companies prioritize the average advertiser higher than a publisher. Verification companies do not provide publishers with scientific tools and solutions that can help them fight fraudulent activities. The only task they do is to uncover the problem.
Publishers have lower chances of successfully negotiating chargebacks with demand players. It’s even more difficult when the results from the verification companies are full of false positives and anomalies.
In order to understand how upside down the industry has become, I collected several examples:
A Rigid Approach: I received a request from a potential client who had a premium portfolio of publishers. He suffered from growing false positive rates on his traffic. The demand players who, back then, worked with aggressive verification policies, labeled and blocked the traffic sources without understanding them. The product itself lacked a flexibility level which advanced customization products provide. We discovered that the product’s logic favored the protection of the advertiser’s side. We assembled a new set of rules that protected the publisher’s traffic sources prior to getting blocked on the demand side.
Complex Relations: Following a successful implementation of our product and pilot test with a video company, I had a meeting with a CEO who informed me that even though our results were great, they would not be able to use our product unless our results correlated with 90% of the client's own results. The CEO admitted that they didn’t really care about fighting fraud activities. They just wanted to avoid exposing their clients to major risks. Many companies don’t have any intention to fight fraudulent activity. As long as no one catches you, you are good to go.
Go Cheap, Lose Money: Never buy a product for marketing or business development reasons. There are many companies out there and the majority of them aren’t fit for purpose. They lack the required flexibility, customization technologies and domain knowledge. Real knowledge. Companies use products and services that promise to clean and eliminate all risks but in reality, the products are full of technical errors and false positives. For example, you might be able to review flagged traffic, but you will never understand which part of the traffic triggered the system - client reports might present fraud categories vaguely. Verification companies which don’t share full reports with their clients are simply a waste of money! They might open new opportunities thanks to having their hallmarked rubber stamp on your website, but in reality, you will be left without adequate protection.
The Sampling Phenomenon: verification companies have realized that publishers, advertisers and trading platforms like to be associated with the reputation of their security products. The verification companies will offer a sample package that will scan only 2%-10% of your traffic and base their strategy around this.
Brand Safety: Many verification companies don’t offer sufficient flexibility in their products. A good example is when publishers contact their vendors due to unusual high fraud rates on their traffic, triggered by certain keywords (such as CORONA) in their content which causes automatic blocks and a decrease in revenue.
Accreditation Companies Are Not The Solution,
They Are The Problem.
The online ad industry is based mostly in the U.S. The natural assumption was that the U.S. government would facilitate high-bar standards and rules. When the government didn’t fill that void, accreditation companies stepped in using their own policies and standards. They focused on models to examine and evaluate work methods but rarely checked the quality of the products themselves. It’s much easier for them to tread lightly and avoid dealing with quality standards. This way, they can sidestep responsibility. The problem with accreditation companies like MRC (Media Rating Council) or TAG (Trustworthy Accountability Group) is that the application process is expensive, overly long and lacks transparency. In addition, once you are a member, you can choose who will become a member. Do you understand the problem? TAG promotes three levels of membership. Why is that? If they are a real accreditation company, shouldn’t they have one standard for all?
MRC’s evaluation process costs over $100,000 and hundreds of man hours across departments and teams. Ernst & Young monitor the process. Having worked with accredited products and companies, I witnessed at first hand how we policies that often didn’t help and processes that few people being created. The goal with these companies is the status and not the process.
In my opinion, these companies will never be neutral. Publishers should be wary. Organizations which force or promote private regulation are dangerous. The private accreditation companies often don’t know the latest fraud strategies, nor the latest technologies and techniques to fight them. They care more about the work process that led to the detection but, crucially, not how you detected the problem.
Instead, put your trust in ad fraud professionals and their in-depth, first hand knowledge.
As an ad fraud professional with many years experience, here are my key tips:
Set a working method with your advertisers – expectations, whitelists and blacklists, Inventory and quality and deadline to receive feedback.
Ask your advertisers and partners what feedback they can give in advance and when you can expect to receive this feedback.
You know your traffic better than anyone else. Don't promise something you can't provide. It's that simple and can save you problems in the future.
Partner with companies which provide full reports in real-time.
Choose verification products that update regularly and collaborate with you to protect your business. How do you know? Ask companies that already work with different products and get reports. If you don’t have anyone you can trust, contact me or sign up for my next webinar on an introduction to verification products.
Review the T&Cs of each company and make sure you don't have unbalanced terms.
Set a ‘Crisis Process’ with your partners to make sure that you don't lose the relationship.
Check that your fraud verification product is flexible and can adapt to the industry on a weekly basis. You can ask for a proof.
Ask yourself if you really need external products? You can create internal logic that will work just fine!
Never buy cheap products (see point no. 3)
If you do want to test/screen a sample of your traffic, make sure to sample relevant traffic rather than just waste all the tests on the first traffic you receive.
Consider accreditation products as part of your business development goals. These organizations have endless powers. However, a membership to their clubs will not necessarily save you from fraudulent activities, but it will nevertheless open many doors.
About: Yehonatan Reut has almost a decade’s experience in tackling ad fraud. In 2011, he led the anti-fraud efforts of Matomy Media Group. In 2015 he joined Protected Media, building their customer support team, creating their integration processes, and opening their office in New York. During this time, he worked on nearly all of the various verification products on the market. Since then he has worked with accounts such as for Taboola, RevContent, Facebook, SpringSer, Smaato and others. Yehonatan is an ad fraud specialist.