In the midst of the COVID-19 pandemic and its devastating effects on the global economy, many companies are monitoring operational efficiency more closely. Likewise, growth marketers are analyzing campaign performance with greater scrutiny.

Like many others, we’ve been looking closer at our data to maximize ROI for our partners. During this process, we uncovered an alarming trend: It appears that mobile re-engagement campaigns are particularly vulnerable to ad fraud. In order to better serve our partners and the entire mobile industry, we’re sharing our insights.

How does re-engagement fraud work?

As mobile marketers know, ad fraud, while unfortunate, is nothing new. It’s been around as long as the mobile industry itself. However, conversations concerning fraud often center around user acquisition campaigns, app spoofing, and click injections. As we’ve discovered, re-engagement campaigns are also vulnerable in ways you might not expect.

Fraudsters are constantly looking for opportunities to steal or manipulate advertisers’ organic results. There are a few ways this works with re-engagement:

• Fingerprint attribution. Since mobile app re-engagement marketing is based on certain identifiers for iOS and Android devices (IDFA and ADID, respectively), the majority of clicks must also use those same identifiers. However, in cases of loss during transmission, attribution is kept open by the fingerprint method. By taking advantage of fingerprint attribution, fraudsters can co-opt attribution of organic performance by generating a large number of click signals without IDFA/ADID.
• Click spamming. Media creators typically receive device IDs from advertisers for targeting purposes. Fraudsters can then take those IDs and send impressions and clicks that didn’t actually occur, resulting in false performance metrics.

Identifying problematic partners & fighting back

While investigating the data, we also figured out a way to fight back against re-engagement fraud by identifying dishonest parties. How? Take a look at the graph below — it represents a real case that we analyzed.

The horizontal axis represents the time after clicking and the vertical axis represents the number of first purchases that occurred at the time. Using this information, we analyzed click-to-action time. As you can see, while most actions for Moloco(Network E) and Network D occur right after the click — actions decrease rapidly as time passes (which is a very typical actions distribution). Networks B and C see their action increase as time reaches around 60 minutes subsequent to their click. It seems that for this advertiser, the fingerprint setting was set to 60 minutes and that Networks B and C have been acquiring actions by exploiting the fingerprint. As with UA campaigns, it’s possible to identify whether your partner is generating false clicks by checking the basic CTR and click-to-conversion ratios.

So what can you do? The first and easiest measure you can take is to work with a trusted partner with proven fraud prevention methods. A good MMP can connect you with a trustworthy ad network and keep your performance safe and accurate.

Need a list of partners now? Check out AppsFlyer’s Performance Index or Kochava’s Traffic Index.

There’s a second step you can take as well. For live campaigns, you need to sweep for fraud in greater detail by having media sources share impression logs, IDFA, and the publisher’s data on impressions and clicks. We recommend keeping the fingerprint turned off whenever possible, but if you need to use it, your MMP should keep an eye on how much conversion occurs through fingerprints. By doing so, you can prevent fingerprint abuse and extract CTR and CVR information for each source in detail.

If you come across some media sources with suspicious data, suspend your collaboration with that partner immediately. Just one fraud channel can disturb the entire marketing performance and hinder you from making reasonable decisions based on accurate data.

Lastly, on top of having ROAS indicators for each channel, we recommend measuring the actual performance of each channel by conducting an incrementality test. However, the cost for testing and the efforts required to eliminate test noises should be considered before making that decision.

We’re Here to Help

To fight fraud in the mobile environment and provide precise ad fraud analysis, Moloco has created an invaluable tool called DoubleCheck. This proprietary service measures true install quality and protects advertisers from all types of ad fraud, including the ones commonly seen in re-engagement campaigns.

If you want to eliminate re-engagement fraud and achieve high performance, please contact us through our official website or reach out to my LinkedIn account. We will keep your data clean and ensure adherence to the highest standards all the way through.

Thank you,

Hyunchai Lee