If you have ever placed an ad within an ad network, chances are you know only too well what fraud is. Every half a year major research companies publish reports of advertisers’ ad fraud losses and promise that you will lose even more money next year as if this was only getting worse.
On the other hand, every now and then there is a network that claims they managed to eliminate fake clicks/impressions/installs once and forever. What is closer to the truth and do you really have any chances to avoid paying for the things you didn’t intend to?
(Yes, fraud is not all about clever bots watching your videos and liking them: we can think of a situation when this is exactly what you expected you happen, the situation changes when Ugandan likes turn out to be from Gambian devices)
The answer couldn’t be simpler: this is a never-ending battle. Don’t trust sweet voices telling you they don’t make mistakes. The best minds on both sides strive for inventing more and more sophisticated ways to be one step ahead of each other.
Still you, as an advertiser, are the first person to save yourself from throwing your budget down the drain, and this is all about both knowing your own product well and making sure you choose trustworthy agencies or publishers. Even if you are lucky, constant vigilance is the best way to protect yourself. We have created a set of golden rules you should follow to prevent dealing with traffic suppliers after the milk has been spilled:
- Know your standard performance metrics. Keep track of your standard CTRs, the conversion rate for each traffic source, the more information you have now, the stronger your position in disputes to come is. For example, there are situations when your non-organic purchase rate is higher than the average. That warms your heart and nothing seems strange until you look at the number of clicks. The publisher might say that they just have cheap impressions in bulk, but most likely they use click flooding to befool the last-click conversion attribution.
- Second, keep track of your best-performing traffic sources. A white list can save you where blacklists fail. Set transparent and hard-to fake KPIs. Do not hesitate to ask about additional/soft KPIs. If you think your CPA is the only thing you care about, think twice, or you might be buying your own organic users’ purchases from someone else (see 1.)
- Monitor your campaign activity or ask the traffic provider to do this. You don’t want your nice and neat private banking solution ad to be run on adult traffic even though you don’t pay for that in the end. You just don’t want that happening. Haste makes waste, so choose an agency that can auto-stop the weirdly performing creative even during the World Cup final.
- When testing a new publisher/agency, think of some non-performance-related targeting settings and change them. For example, you own an Android dating app and you are still growing your user base without earning from the app directly. Most likely, you don’t care about what devices your login with as long as the app doesn’t crash. Choosing random 100 device models (out of the 1000 most popular ones) every week would make it harder for a botnet to fake a specific list precisely.
- If, by any chance, you don’t have a direct money-related metric to base your KPIs upon, choose your affordable fraud level (=the amount of weird/badly performing traffic you can still pay for). Strange as it seems, this is perfectly reasonable if you need huge volumes. Whitelists, strict targeting, and well-weighed KPIs would definitely save you from many of the evils of the ad world, but, sadly, this often comes with your audience/reach shrinking to zero. When everyone seems to have your app installed and your executive board still wants growth, calculating and taking risks can be beneficial.
- Choose less complicated ecosystems and try to buy as directly as it is comfortable for you. The concept of RTB is awesome and some of its realizations are extremely good, but the more participants the chain has, the worse is the performance impact, both in tech and financial terms. Middlemen make your ads more expensive, endless “black boxes” and scripts your ads are run with make it hard to define the real placement/publisher.
- Use a third-party fraud prevention system. A problem shared is a problem halved (sadly, this is vice versa for your budget, but still worth it)
What about new technologies? There is a bunch of ad tech market leaders responsible for innovation. Not surprisingly, they only reveal the basic principles of how they tell a real action from a fraudulent one because this includes high-level programming, math models and neural science. Even if they were open-source, the aimed-at clockwork fraud protection mechanisms would not be that easy to copy, implement and maintain.
This is equally true for blockchain solutions. We have found some 10+ ad tech companies who put their trust in distributed databases and smart contracts. Most of them seem pretty legit, some even have decent chances to grab their market share (especially when they work with finance fraud too) but all of them cannot but use the same fraud detection logic as the regular RTBs. They all have to create algorithms to decide whether something looks natural or not, the only difference is that it is far more complicated to hide the unwanted data or set your partner up in terms of payments.
What does it all mean? “Constant vigilance!”, we would repeat. Following the 8 abovementioned fraud prevention steps (actually, there are many more) is a challenge, so leaving it to those who can take care of everything and more is only natural.