Ad fraud is a huge industry topic; but mobile ad fraud receives comparatively limited airtime. How big an issue is it and what is being done to tackle it? Ilkka Lassila (pictured below), CTO, Widespace, tells ExchangeWire what his take is on the growing problem.

ExchangeWire: When we talk about mobile ad fraud, what are we referring to specifically?

Ilkka Lassila: Although a majority of criminal activity occurs on larger devices, the phrase “wait five minutes and it’s all mobile” is one that’s becoming more commonplace, when it comes to ad fraud.

To understand the phenomena, we need to look at different business models; that is, we need to understand the ways to make money in this increasingly complex ecosystem.

We see different players packaging publisher inventories to, arguably more valuable, advertising units. These units are then sold to marketers who do everything they can to measure the effect. A major slice of this business is bulk impressions and clicks arbitrage. Most of the technology used is not designed for mobile, and even less designed to address possible fraud. This, then, opens up lucrative opportunities to fraudsters around the globe.

There are many numbers on the cost of mobile ad fraud to the advertising industry – what figures are you seeing? Is the issue as serious as reported?

Fraud levels across mobile are still relatively low, compared to desktop. We have not seen many trustworthy studies about mobile fraud levels; but, in total, the level of invalid traffic is significant. When talking about lost money, it’s hard to differentiate between clear criminal activity and poor quality. All in all, there is a huge amount of money wasted in bad-quality traffic. It’s easy to generate lots of traffic, with no real advertiser value. And, in the world of programmatic, it's fairly simple to monetise on it as well.

What exists in the marketplace to combat mobile ad fraud? Is it effective?

Controlling and monitoring self-service channels and disabling publisher self-sign-ups proved to be great mechanisms to promote better values for us at Widespace. While it was a step away from automation, it was a leap forward in fraud prevention.

Even with all the latest technologies to analyse ad fraud, a few more traditional methods seem to be the most effective. In every hacker’s handbook, the critical link is typically always human. Expensive technologies can crack almost any system temporarily. However, more affordable, meaning potentially more profitable, ad-fraud setups rely on identifying weak spots or bad human choices. By knowing who you work with, and knowing the market you are in, you can avoid a majority of the issues. Choosing publishers who are in the business for the long term also offers higher security. It is in their best interest to grow their user base and inventory organically, ensuring quality and auditability.

Investigating volumes, market shares, inventory growth and variation will present an idea of publisher actions. Monitoring ad placements, comparing results, and evaluating load times, viewability, and other metrics, will create a stronger base for quality. Whenever anomalies or alarming patterns are flagged, a human counterpart to call to, to ask for more data and to double check your end-to-end figures, is invaluable. While it may not be an ideal model for each marketer in such an automated and programmatic world, it creates trust, transparency, and quality. Choose a partner you trust.

Where lies the responsibility in the ecosystem for tackling mobile ad fraud?

We need better quality standards and guarantees to replace bulk impressions and clicks. Better quality guarantees can be mapped to more direct advertiser value. When it comes to branding, the focus should be placed on brand awareness, interest or consideration, instead of artificial clicks and attribution funnel.

In the last six months alone we have seen great initiatives in the industry to increase quality; IAB LEAN is a proposal to deliver better and more user-friendly advertising in general. Google AMP offers an optimised HTML-based stack to do that. IAB DEAL proposes an industry-wide pattern to negotiate with Ad Blocking Users. All in all, these initiatives strive for better quality, counter both fraud and zero-valued ads. However, we need more of these to tackle mobile ad fraud specifically.

A majority of anti-fraud companies currently rely purely on technical prevention mechanisms; which are important for both automated quality control and reports for discussion points between different partners. In the simplest form of prevention, a delivery system should verify the device, user agent, domain, ISP, frequency of ad requests, in-screen rates, interaction rates, etc. – and compare all of this to known human behaviour patterns.

It is the ad-serving side's responsibility (ad server, DSP, DMP) to identify and know the user. At Widespace, the way we down-prioritise short-lived users has proved to be a very effective way to prevent fraud. Large amounts of fraudulent traffic creates new users. This bypasses frequency capping, and delivers increased reach to greedy publishers.

From an advanced delivery point of view, frequency control requires user history and a profile for added accuracy and higher quality.

A distinct challenge with fraudulent mobile impressions is their low CPMs, relative to non-fraudulent impressions. How can marketers ensure they are getting the value back on their bottom line when buying above-board inventory?

Mapping programmatic marketing efforts to offline data is the best metric out there, at the moment. However, simple common sense still applies: if it’s too good (cheap) to be true, it probably is. Marketers should use advanced analytics, make high transparency a requirement, and map the raw marketing data to their business figures to double check.

Bots are becoming more intelligent – is mobile ad fraud just an evolving issue that can't be eliminated? What does the future hold?

To operate a business in digital advertising and address mobile ad fraud, there is one simple rule: be transparent with your partners. Offering access to raw data allows clients, and partners, to evaluate and build up knowledge of transactions. Human-to-human trust is such an important factor.

By being transparent, you can add reasonable disclaimers on your quality guarantees, and allow each partner to build up their own knowledge. The ones with high quality can always afford higher levels of transparency.

In the future, we will start to see more and more high-level quality standards that go beyond impressions and clicks. This includes both the protocols used in standard integrations, as well as currencies charged from clients. Digital advertising will become more and more value based.