Partner marketing can be extremely valuable, but it can also leave brands vulnerable, says Neil Ranatunga. Protecting your business from fraud doesn’t have to be difficult, here are four sensible tips every marketer should know.

Fraud is a fact of life in Australian digital marketing today, though it affects different categories in different ways. In impression-based media buying, for example, most Australian marketers know that there are lots of bad actors trying to steal from their brands. By contrast, because partner and affiliate marketing revolve around verified purchases rather than more game-able metrics like impressions and clicks, the incidence of fraud is lower.

That said, Australian marketers are naturally interested in getting ahead of the challenge before it significantly affects their businesses. If you want to start assessing your partnerships for fraud and preventing this form of theft from hurting your results, here are four simple tips:

1. Don’t panic, but pay attention

Fraud rates are not at crisis stage in Australian partnerships. But any form of fraud is theft. Just as you must take sensible steps to protect yourself from pickpockets and other thieves, so too should you be aware and respond to threats of fraud on your business.

One of the best ways to identify possible fraud in your marketing is to review a set of key metrics, noting any unusual or peculiar data. Many of the most common – and costly – forms of digital fraud can be detected through relatively straightforward data analysis.

While peculiar data does not prove that there is fraud in your ecosystem, strange figures and trends can serve as road signs pointing brands to areas of concern. Better still, real data provides the sort of concrete evidence you need to prove that investigation is necessary. Suspicions are one thing. Facts are quite another.

2. Investigate variances

Fraudsters want to maximise the amount of money they make with a scam. That means that they usually go for scale. While successful large-scale fraud can cost a lot of money, it can also be easier to spot in data.

Look for big variances in your most important metrics versus the norms you expect. For most partner marketers, the desired action is a purchase, so purchase metrics warrant the most scrutiny. If your program is more focused on traffic driving, take a close look at your click metrics.

Here’s what this process looks like, using publisher conversion rates as an example. No publisher converts all of the traffic that it drives to a brand. Rather, a subset or fraction of users will convert while others will leave without buying. When an abnormally low or high percentage of visitors driven by a publisher make a purchase, marketers may wish to dig further:

• Abnormally low publisher conversion rates can be a sign of click fraud, in which an advertiser simulates clicks to a publisher’s site. Naturally, this tactic is most commonly used in CPC campaigns. But you might also see peculiarities, particularly if you pay for prominent features on partner sites.
• Abnormally high publisher conversion rates can be a sign of creative fraud, in which your ads have somehow been changed to drive greater buyer interest. For example, a fraudster could advertise a highly appealing but unauthorised offer. Customers buy, but then are disappointed with you for not making good on the deal.

There’s no magic partner conversion rate and publishers will naturally differ in conversion rates. When you spot a publisher with an extremely high or low metric relative to the normal range across your partners, talk to them to understand why. There may be very legitimate reasons for the disparity, but ask the questions and collaborate to protect your business.

3. Be vigilant

Big differences in existing results are a great way of spotting things to investigate, so too are sudden changes in your marketing performance. Anomaly detection refers to examining data for peculiar changes that may indicate a problem such as fraud.

It’s important to remember that metrics around any marketing program do fluctuate. But there is likely a normal range of variation. When changes go beyond that range, you’ll want to investigate further. By combining anomaly detection with a powerful alerts system, you can automatically be apprised of data changes that warrant further attention.

Some anomaly detection systems are manually configured so the user sets the range beyond which they will receive alerts. But an increasing number of tools use machine learning and artificial intelligence to identify those normal ranges and automatically determine which changes warrant further scrutiny. Such systems are usually better and far easier to use. Computers are often superior at identifying significant changes in data than people and AI (artificial intelligence) helps eliminate time lost to the trial-and-error process.

4. Having an open dialogue

Extreme differences and changes in data are reasons for investigation, not accusation. Shared data and analysis are a great way to collaborate with great partners for even greater results. Work with your data provider to understand what data anomalies are worthy of investigative time and attention.

For example, some types of partners will naturally have much higher or lower conversion rates because of their traffic composition and role in the customer journey.  By creating direct relationships with partners, you can glean best practices for working with them and identify your more engaged and trustworthy partners. Honest sharing can also help partners spot fraud they might not yet know about in their ecosystems. Direct communication can help you both address issues and grow together.

The world of partner marketing is constantly changing. Many of those changes are for the better. Some aren’t. But these sensible tips can help improve your results and protect your business from outside threats.

Neil Ranatunga is strategic partnerships director at Performance Horizon.