Ad fraud is an expensive global phenomenon. With projected losses of $5.8 billion USD globally for the 2019 financial year, it is evident that this style of criminal activity is extremely lucrative for technically competent, corrupt organisations. And make no mistake – these are exceptionally well organised, well set up systems. Nowhere has been harder hit than Asia Pacific countries (APAC), and Southeast Asia (SEA) in particular, with regard to the use of ad bots, malware and digital farms. 

According to a recent report by AppsFlyer, the level of financial exposure to FinTech and e-commerce enterprises reached a staggering $650 million USD in APAC for the six months between November 2018-April 2019. The SEA region has been exposed to the tune of $260 million UDS, with India following at $188 million USD, and Japan-Korea with nearly $100 million USD each. 

The “We Are Social Report” in 2018 claimed that in Indonesia, Malaysia, Philippines, Singapore, Thailand and Vietnam there were 391 million mobile internet users. Given even conservative predictions based on those previous year-on-year increases, the number will increase substantially by this time in 2019.

Can AI and machine learning out compete hackers, fraudsters, and the dark web?

Ad fraud focus

Ad fraud is the process of using the viruses, software hacks – like click hijacking, bots, click farms and any other way to derive revenue by providing a ‘human-like’ appearance for a variety of actions. Artificial viewing of an ad, fake download of an app, fraudulent registration of information are all ways which are used to siphon money away from a region with no economic return.

Fraudsters face limited risk and a lot to gain for little effort. The bots do it all, and while each individual action made by the bots is worth very little, when this is multiplied in the way that ad traffic is streamed, it amounts to a significant amount of money. Substantial enough to be of interest to major crime outfits and more lucrative and less hazardous than many other avenues. 

Previously, there were actual physical ‘device farms’ complete with buildings full of computers clicking, infiltrating apps, registering fake accounts, install hijacking, and ‘viewing’ ads for the prescribed amount of screen time (2 consecutive seconds) in order to receive remittance. In some unautomated cases ‘low paid’ workers – human trafficking instantly springs to mind – were involved in the click process. The current trend for legions of bots is a more straightforward, more productive, and more cost-effective means of infiltrating apps or ads and committing fraud for professional fraudsters.

man sitting facing monitor

The United Nations 2010 report “The Globalisation of Crime: A Transnational Organised Crime Threat Assessment” recognised the problems which cybercrime causes, of which ad fraud is a significant part. Times have changed since then and the damage to each region’s governance and stability as well as the potential to undermine social trust, and it’s all important economic growth. It is even more profound. Global dependence on all things tech has provided an endless opportunity for nefarious systems to interrupt the right flow of information for their own financial gain. 

Understandably The World Federation of Advertising is concerned that with the current rate of market growth, associated fraud and the projections for both, this type of corrupt behaviour is set to become the 2nd biggest organised crime enterprise, after drug-related activity by 2025.

Emerging FinTech trends in Southeast Asia

The AppsFlyer report regarding the depth of the ad fraud issue in the APAC region claims that approximately 25 out of every 100 non-organic installs in the region are fraudulent with finance, shopping, and travel being the top 3 most targeted e-services. The scale and sophisticated nature of app fraud makes it harder to detect and manage. It is also under-reported, with finance apps reaching the highest fraud rates across regions, up to 2.5% higher than other app types, while gaming apps retain the lowest rates during the assessed period. 

Growth culture

Within APAC, there are substantial differences in the figures of disclosed fraud between the regions due to the varied economic and social trends at play. Australia’s ad fraud figures are markedly lower (peaking at 9% in the assessed period) than the highest in the SEA region, Vietnam (peaking at 50%), in spite of them both retaining positions in the APAC group of countries.

Rapidly increasing internet penetration and steadily growing digital behaviour habits are contributing to the issue. Currently, the figures range from an average screen time of 3.6hrs/day across Southeast Asian nations, with Thailand topping the global charts with 4.2hrs/day. When you mix these statistics with high market demand for eCommerce, underdeveloped app security, and often a Cost Per Install (CPI) business model, the culminating factors of the problem is evident. 

CPI doesn’t necessitate an ‘action’ and is therefore easily replicable or hijackable by the bots, generating an enormous revenue from an already high volume of fraudulent traffic in the localised networks. All of these are the components of the perfect, fraud-laden, storm.

Forward planning

The AppFlyer report and others in its field are undoubtedly concerning. It highlights the need for targeted resources in app development and online security to provide on-going protection for customer and business alike.  

The report, while clear in its dissemination of a genuine issue, has been produced by one of the big players in the market of app software protection and therefore has a vested interest in driving interest and finance into this arena. It is important to note that the increase in the issue of ad fraud is creating a booming business in ad protection, with a market growing symbiotically to its antithesis.

Ashish Shah, Founder and CEO of Vertoz, writes that blockchain, with its capacity for transparency and a static ledger system, might be able to “provide [a] trustworthy ad-tech solution.” If Shah is correct, blockchain could provide the necessary level of end-to-end security, trust, accountability, and transparency, that programmers and providers could use while also retaining a degree of autonomy. What appears more likely is an AI machine learning solution which can manage, sort and perhaps ‘clean’ big data, fast.

The issue for SEA will continue to be the increase in eCommerce and FinTech platforms and their built-in security. It will take a big push to get ahead of such organised crime, and then an equally big leap to put sufficient structures in place to ensure the longevity of economic growth that the region can potentially reach. 

11 percent decline in two years is particularly impressive considering that digital ad spending increased by 25.4 percent between 2017 and 2019.” While industry leaders are congratulating themselves on this, the impact of the global distribution of ad fraud tells another story.