B&T goes one-on-one with InMobi’s senior vice president of global brand and programmatic, Anne Frisbie, to talk all about the threat that is ad fraud, and what brands and marketers can do to prevent it.

What are the common ad fraud issues on mobile in the Asia-Pacific region, and in particular, Australia?

There are two common fraud based issues in Australia: install-based fraud and misattribution.

Install-based fraud occurs when humans, bots or scripts initiate a fake installation process, or when a user’s app-install behaviour is imitated on their device. This type of fraud often goes undetected as the clicks, installs and users are non-existent.

Misattribution happens when installs are valid but credit is stolen from clean networks. Unlike the desktop world, mobile ad fraud spawns and thrives on the opportunities arising from the lengthy duration of the mobile journey from impression to install to transaction. Malicious players cripple networks that provide ‘fraud-free traffic’ with the influx of misleading data, which acts as both the input for campaign optimisations as well as a measure of network performance. There are three main forms of misattribution:

• Click cramming, where fraudsters target a certain set of device IDs known to engage with in-app campaigns, driving timely clicks from these device IDs to steal attribution from other networks. This further increases ad spends onto malicious publishers, further damaging campaign data.
• Ad stacking, which occurs when multiple ads are layered under a single, completely visible ad. Any click on the visible ad triggers clicks for all the other ads stacked underneath. Accidental clicks are also quite common. This is where an ad is disguised as an arrow button that appears to lead to the next page of an article.
• Click sniping, where a fraudster places a click right before the app-open in order to claim credit for the conversion.

Is mobile ad fraud going to go away anytime soon?

It will take some time for advertisers to become aware of the capabilities and limitations of a partner’s fraud detection algorithms, as they directly correlate to campaign performance, budget allocation and ultimately, revenues.

Trackers can help eliminate ad fraud by measuring campaign metrics using mean time to detect (MTTD) instead of mean time to interrupt (MTTI), and by sharing their fraudster blacklists globally.

This requires a coordinated effort and suggests that ad fraud will be with us for some time to come.

What are your top tips on how brands and marketers can prevent ad fraud on mobile?

Advertisers concerned with ad fraud need to partner with networks that are investing heavily in detecting install-based fraud. When it comes to misattribution, brands need to educate themselves on the limitations of the data they’re presented with. Becoming familiar with the nuances of fraudulent and suspicious behaviour is crucial in analysing irregularities in campaign performance.

With this in mind, there are three best practices to creating a successful in-app mobile performance strategy:

• Partner with a third-party mobile tracking and in-app attribution partner with the correct certification requirements.
• Require impression (not just click-level) transparency for all buys, whether you are paying for cost per impression (CPM), cost per click (CPC) or cost per action (CPA). Impression level visibility is critical for better understanding true incrementality of your direct response media spend.
• Don’t solely rely on last click attribution models that are increasingly understood as flawed, and encourage fraudulent behaviours by bad actors. Instead, incorporate view through attribution (even if a small percentage, and for a short window) to all media partners equally. View through attribution will also help enable you to enforce impression level visibility with your media partners. Additionally, you may want to consider mid-funnel attribution such as video, which completes to a near equal level as the last-click attribution, in order to better account for the value of video advertising.

What strategies should brands and marketers be implementing to measure ad fraud?

Education is the best strategy brands can undertake when it comes to protecting their campaigns against ad fraud. By educating themselves on the capabilities and shortcomings of their mobile attribution partner’s fraud detection technology, brands are better able to understand the success of a campaign measured against legitimate performance, budget allocation and revenues.

So, becoming familiar with fraudulent and suspicious behaviour in mobile advertising becomes a crucial means of identifying and understanding irregularities in a campaign. This then allows marketers to work together with their attribution partners to collaboratively ensure a genuine and successful campaign, free of ad fraud.