Practical tips for beating ad fraud
If ad fraud is so easy and lucrative for criminals, what can advertisers do about it? Tom Evans and James McDonald share their advice to help you minimise the impact of ad fraud on your ad spend.
Imagine buying airtime on a TV station that doesn’t broadcast to anyone or buying space in a newspaper that doesn’t get printed or distributed. You wouldn’t pay to advertise to no one, would you? And yet part of your online spend is for space that no one sees.
A few years back, the World Federation of Advertisers reported that ad fraud was second only to the drugs trade as a source of income for organised crime. Since then, Juniper Research has forecasted that advertisers’ total loss to fraud will rise to $100 billion by 2023, driven by increasingly sophisticated techniques these crafty fraudsters continue to develop to grab as many ad dollars as they can.
We can’t promise to stop ad fraud in its tracks, but we can help you minimise it.
We didn’t need HP’s Business of Hacking report to tell us that ad fraud is lucrative and attractive because it’s so easy.
Let’s imagine a simple ad fraud scenario. They set up a website and scrape some content from a couple of other sites. Next, they set up some banner and video ad units. They buy some cheap traffic from wherever they can, then sign up to a supply-side platform and start accepting bids for that cheap traffic they just bought. Whenever they serve an ad, they make sure that each ad unit has 10 ads stacked on top of each other. Sure, if the advertiser checked they would see this, but they won’t check and the fraudsters count on that. The advertiser will rely on their ad fraud/verification pixels to fire. Which they will. Then the fraudsters will unleash their bots. Every ad served will generate a robot click-through to the advertiser’s site.
The advertisers are happy, counting all the relatively cheap clicks they are getting from their campaigns and plan to spend even more with this site.
The fraudsters are very happy now. They’ve paid very little for some traffic, likely buying bot traffic from their fraudster mates. They’ve sold that traffic – ten times over, in fact, because they served ads overlayed on one another – to advertisers. Each impression has artificially created a bot click and web traffic for the relevant client. The fraudsters have just made a heap of money for very little effort or investment.
A fine impression of an impression
A very large amount of online ad spend is being soaked up by these fraudsters. And in some countries, it’s not even illegal to do this. After all, contractually speaking, the ad is meant to serve an impression and it’s doing that. Unless the contract states that the impression must be human, the ad is fulfilling the contract.
Your media agency is also legitimately making money from this ‘not so legitimate’ transaction. They get paid even though no value is actually delivered. To be fair, they’ve done their best. They haven’t set out to buy invalid traffic on your behalf. They have used a reputable ad fraud and verification company to prevent this from happening. It’s just that the fraudsters are a step ahead of the fraud catchers.
Four tips to beat the fraudsters
Right about now you must be getting pretty angry, thinking about just how much of your ad budget could be going up in smoke. So, what can you do?
1. Conversion events
Create conversion events on your website that would be hard for a robot to spoof. Then ensure your ad buys are optimising to these events. You will still buy fraud, but the fraudulent sites will quickly show as being ineffective against a human conversion point. Make a point to weed out these ineffective sites, or ask your agency to show you that they are.
2. Measure for human behaviour
Find a leading indicator that measures human behavior and ensure that you can link all of your advertising activity, but especially online, back to this measure. Develop measurements that will help you identify issues with the effectiveness of your online buys.
For example, if you are buying online media to drive brand awareness (and we would suggest this is a poor metric to which to optimise, but that’s a debate for another day), then ensure you have a measurement in place for it. We use our own ‘Share of Saliency’ model. This model is a leading indicator for future sales. It is derived from human observed actions and we track the rise and fall of this index against our clients’ media mix. In this way, we can immediately see any effects our display and video advertising are having on our share of saliency. If we are spending online and seeing no positive movement, then we have identified there is a problem in the execution of online buys.
3. Insist on a whitelist
Review all of your media placements and see which ones are generating conversions. Be especially wary of sites that deliver high traffic and low conversions. Build your own whitelist from these results. We’ve done it for all of our agency’s clients and maintain an agency whitelist.
4. Reputation matters
Do deals with premium publishers and walled gardens reputed for low incidence of ad fraud.
Or hey, let’s make a quick buck on building an ad fraud network*. What could we call it …?
* This is a joke. Jokes are funny. Fraud isn’t.
James McDonald and Tom Evans are the co-founders of independent media agency Audience Group.