This article originally appeared in Digital Market Asia.
Whether you’re an all-in-one marketer at a startup, or a digital specialist at an established global brand, chances are you’ve earmarked a good chunk of budget towards digital advertising. According to a recent TrafficGuard whitepaper, US$130 billion will be spent on mobile advertising this year alone, accounting for 60 per cent of digital advertising spend.
Digital advertising’s appeal is easy to see – global time spend has boomed over the past decade, offering advertisers a channel to connect directly and immediately to their intended audience. The unparalleled degree of customisability, as well as (relative) accuracy of tracking and measuring has been a key driver as well.
However, for all the value it offers marketers, the rapid growth of digital advertising has instigated a catch-22 situation. The increasing lucrativeness of the industry has fuelled criminal ingenuity, with fraudulent practices constantly mutating to avoid detection, jeopardising return on ad spend (ROAS). Indeed, the same TrafficGuard whitepaper suggests that mobile ad fraud alone will cost brands US$34 billion in 2018 – more than a quarter of total spend. More than half (US$19 billion) of that will be lost in APAC.
Cautionary tales are aplenty. In November, Google and the US Department of Justice released details of the 3ve botnet scam – an elaborate scheme that defrauded advertisers out of US$36 million over four years.
The challenge is that while fraudsters’ tactics are evolving, our defenses are scrambling. All too often, marketers think of the problem in terms of wasted spend – and consequently focus on detecting fraud to pull funds back as opposed to preventing it from occurring in the first place.
Know your enemy
The reality is that fraud today is a far bigger beast than any one party can handle, with various factors that make it tricky to detect. The exponential growth of digital advertising over the past decade has led to a number of new intermediaries – such as SSPs, exchanges, DSPs, and Ad Networks – adding multiple layers between each click and an ad’s ultimate destination.
Meanwhile, on a more localised level, the sheer size of the market in APAC, coupled with increasing smartphone proliferation in emerging nations and high user engagement make it the perfect target.
In such a complex landscape, it should therefore come as no surprise that the impact of fraud is equally complicated. Though often overlooked, additional costs very often include:
- Investment in Fraud Inflated Sources: A lag in fraud detection may mean you unwittingly divert spend from genuinely higher-performing traffic sources to ones that are seemingly helping you to hit your KPIs
- Wasted time: Advertisers that can detect, but not block invalid traffic spend a lot of time interpreting reports before taking action, then reconciling media volumes with intermediaries to try to claw back their misplaced ad spend
- Litigation: Unresolved disputes about what was and wasn’t fraud can expose both advertisers and intermediaries to costly and time-consuming litigation cases. Earlier this year, Uber was embroiled in a tussle with mobile advertising agency Fetch after refusing to pay the latter millions of dollars, alleging that the agency had channelled ad spend to fraudulent platforms
- Diminishing Campaign Optimisation: Traffic suppliers can’t optimise campaigns in real-time as only the advertising brand or business receives reports – if these are not shared, campaigns can’t reach their true potential, ultimately resulting in poor campaign performance.
- Fraudsters Funded: Ad networks regularly pay for their traffic sources before they receive payment themselves – it’s the nature of the business. This means that if the advertiser refuses to pay for traffic it is has identified as invalid, the ad network is already out of pocket – and the fraudsters get paid regardless
The way forward
While a good reporting tool might help you reclaim some of your wasted media spend, it’s important to note that the above costs cannot be recovered by reacting to fraud reports.
The challenge is how you proactively prevent fraudulent traffic before it hits your budgets. The temptation is to slash entire fraudulent traffic sources out of the equation from the get-go, but then you run a high risk of cutting out the good along with the bad.
So what should businesses be asking when choosing an ad fraud prevention solution?
- What data do you analyse, and at which stage in the journey?
As it is, most solutions only analyse a single level of invalid traffic (IVT) – for example, at the impression, click or attribution stage. This restricts focus on a small proportion of the advertising journey. To mitigate sophisticated ad fraud effectively, you need to be analysing multiple stages of the advertising journey to ensure that if fraudulent traffic evades one line of defence, it gets caught at the next one. Think about coverage levels too – does your tool monitor only a sample of traffic, or every transaction? You should also consider whether your solution has the ability to evolve – static rules and behaviour signatures will only be able to do so much when met with an unknown scenario. Look out for machine learning and AI capabilities, ensuring solutions can process large and complex data sets to learn and improve over time.
- How much traffic do you block, and when?
Blocking capabilities vary across different solutions. The ability to block in real-time is essential to minimising the loss to fraud that can occur at multiple levels of the advertising chain. However, it’s also important that your solution doesn’t take a blanket approach to blocking. Extreme highs in traffic blocked might indicate the presence of false positives – leading you to block good traffic along with the bad. Just because there was one fraudulent click from a traffic source doesn’t necessarily mean the whole source is bad.
- What does your reporting look like?
The speed at which bad actors modify techniques to evade detection, and the increasing sophistication of fraud tactics renders conventional, rules—based fraud prevention ineffective. Essentially, advertisers (i.e. brands and businesses) need to know what’s happening in as much detail as possible, and in real-time, to be able to combat fraud. However, insights are only useful when they feed into action. Ultimately, if your partners and intermediaries are still paying for fraud, which you then have to reclaim as wasted media spend, the data only takes you so far. It’s therefore important to consider how reports are going to be shared – for example, some solutions offer reporting accessible to all parties. Ad agencies and networks are then equally able to monitor campaign performance, identify invalid traffic sources, and optimise campaigns accordingly.
The last point in particular is important. Mitigating ad fraud doesn’t stop at your solution – with the multiple intermediaries now involved in an ad cycle, it is crucial that all parties are collaborating, and are able to agree on and understand the basis on which each transaction is invalidated.
The order of the day is transparency. Only when sharing reports with partners can we work together to prevent fraud from occurring in the first place, and reduce its impact on the ecosystem to make the opportunity seem less appealing to fraudsters. Without this information, we’re sitting ducks waiting for our spend to be chipped away.
Continuity and consistency is also key. While swapping ad networks regularly is an industry norm, this practice does ultimately restrict performance. By sticking with your partners for the long-term, you can build on the learnings of each campaign and drive continuous performance improvements.
For too long, the industry has been focused on addressing the symptoms (or rather, a symptom – wasted media spend) of fraud – but where’s the cure?
Reclaiming spend is not going to dissuade fraudsters. The only thing that can do that is a concerted industry effort to tackle the challenge, pulling the carpet from underneath fraudsters’ feet. That means advertisers and traffic sources need to have access to the same, and as much information as possible on traffic quality in real-time – not just at the time an invoice is raised. After all, if these sources can optimise fraud out of their supply, the fraudsters don’t get paid – plain and simple. Make it too difficult to justify the fight, and we will win the war.