Global programmatic ad spend is predicted to reach USD$37bn (£30bn) by 2019, up from USD$14.2bn (£11.5bn) in 2015. However, as programmatic budgets boom, so do the opportunities for fraudsters to take advantage. In 2016, fraud and viewability concerns ranked as the biggest barriers facing programmatic adoption. This edition of ExchangeWire’s Now & Next takes a closer look at what needs to be done to stop an industry which is emerging as one of the biggest organised crime enterprises in the world.
The many faces of fraud
Despite being as nascent as the ad tech industry, ad fraud is sophisticated and takes different forms. Two of the most damaging are:
– Ghost sites: Fraudsters create large content farms with whatever topics a marketer may want. They then use bots to make ‘qualified audiences’ appear, with zero publications or viewers actually involved. These are posted to all the ad exchanges, until one unwittingly agrees to list them, and publishers choose to work with them.
– Phony traffic brokers: The publisher wants to increase traffic to their site and goes to a traffic broker site that’s actually run by a fraudster, who promises volumes of highly qualified users. The publisher then pays for a particular package, full of false traffic.
Messing up measurement
The problems for advertisers are evident. If you are investing on a CPM basis, you cannot be sure how many of the impressions you are paying for are from a human and how many are from a bot. This results in advertisers paying for human impressions they never actually received. Conservative estimations predict the total cost of fraud could stand at USD$50bn (£41bn) by 2025.
In addition to the monetary cost, it has caused advertisers to completely rethink their metrics. Assuming all impressions, clicks, or traffic are recorded, if any proportion of these are from bots, their will now be a flaw in the system. As we cannot guarantee what proportion are fraudulent, it is impossible to know how skewed current measurement systems are, and how inaccurate the results are in comparison to reality.
Publishers are by no means immune to the issues caused by ad fraud. Ghost sites that offer fake content and fake audiences are diverting the advertising dollars of unwitting media buyers away from legitimate publishers. What’s more, phony traffic brokers cause advertisers to be suspicious a large portion of their views are from bots. This not only diverts funds away from programmatic buys, but also creates reputational harm for publishers.
On mobile and in-app
Fraudsters are smart – wherever the money is, they won’t be far away. A clear sign that mobile programmatic is coming of age is that fraud practices employed on desktop are now moving to the mobile ecosystem, and a big focus for fraudsters is in-app ads.
Although official app stores provide a layer of oversight to prevent fraudulent apps from entering the marketplace, bots still find their way onto mobile devices, typically from users downloading from third-party app stores or websites, where malicious app developers place sophisticated, legitimate looking apps.
Ad fraud appears on over half (52%) of uncertified apps and in nearly one-tenth (8%) of certified app traffic. While Android is more susceptible to fraud than iOS, the latter is not immune to fraud and is now being targeted heavily given that, if successful, payouts will be far higher than on Android.
But mobile won’t be the final frontier for ad fraud. As the industry continues to advance areas such as programmatic TV and audio, and budgets start to flow towards these categories, fraudsters will re-divert their attention accordingly.
An unstoppable force?
Given the size of the problem posed by ad fraud, it is perhaps surprising that there is no panacea. Attempts to combat the issue so far include the Joint Industry Committee for Web Standards’ certification in the UK, and machine learning to detect potential fraud-driven anomalies. However, it’s widely accepted that more needs to be done to tackle the issue. So, what extra measures are being taken?
– Third-party vendors: Publishers and buyers need to deploy monitoring tools. This will most likely be in the form of a third-party vendor, who will scan creatives a few times every hour, keeping an eye out for suspicious viewing activity. This is largely a matter of scale, as neither publisher nor advertiser has the time to analyse all of their own creative.
– Metadata: Using metadata from the app stores to execute strategic and safe buys is one possible route. Media buyers can leverage this metadata to ensure they only buy inventory on apps with high star ratings, which have over a certain amount of downloads and customer ratings. This signals a high-quality, legitimate app, which, as mentioned earlier, is less likely to be exposed to app-install or -engagement fraud. For this to work, stronger relationships between app stores and media buyers must be established.
– Transparent language: Codes of conduct, such as those adopted in Germany, are another step in the right direction. These require signatories to provide information on their data sources, data collection methodology, and policies. It also demands suppliers to install safety measures against ad fraud and provide the tools to check that fraudulent ads are excluded from the traffic. There also needs to be language placed into RFPs and IOs requiring publishers to identify all third-party sources of traffic.
When will it end?
On the the surface, it can seem that buyers and sellers face a seemingly impossible battle. If brands continue to focus on vanity metrics, such as clicks, when determining campaign success, fraudsters and bots will continue to take advantage of this, following the industry to areas such as programmatic TV and audio. However, it can only be a matter of time before marketers get smart to this fact, and develop more sales-based KPIs, minimising the need for high CPM rates, and thereby disincentivising ad fraud.
However, this will not happen overnight, and the future for ad fraud is still lucrative. As well as programmatic TV and audio, expect fraudsters to delve deeper into mobile apps. With leading brands now seeing huge amounts of sales completed via apps (one-fifth for Starbucks in the US), the opportunity to engage consumers through this method is increasing, and the opportunity for fraud here is rising correspondingly.
There appear to be plenty of solutions to tackle the problem of ad fraud. However, bots will continue to slip through the nets of buyers and sellers worldwide; and the majority of the methods are challenging to implement or will only result in small improvements. It is only a matter of time before stolen budgets force advertisers to ditch vanity metrics and take up more sales-based KPIs.