Recently I had a thought-provoking conversation with a marketing manager, discussing ad fraud. Sitting in a café, it took 75 minutes to open her eyes about its presence in the advertising industry. I decided to transfer the essence of our conversation to a blog post to the benefit of others.
Ad fraud is predicted to grow to $40-140 billion in the coming decade, which means almost all participants of the industry is or will be affected by it. Whether it’s deliberate deceit or a simple misstep, the banner impressions and site visits can be affected in a negative way, resulting in lost revenue. As tough as it may be, ad fraud can be tackled by sophisticated methods to protect the quality of traffic.
By doing that, you can optimize media buys and improve many others aspects of your business:
1. Optimize ROI – your money is spent on ads the right way
2. Segment sources – pinpoint the best and worst traffic partners, eliminating the latter
3. Gain insights – understand exactly what is happening with your ads and visitors
4. Identify non-human traffic – understand your ratios and weed out these impressions and sources (this is now an MRC & ABC standard)
5. Work with real-time alerts – get suggesting actions based on events related to your traffic
6. Make sure your traffic is worth its price
There are many techniques for reaching these targets via ad fraud prevention. You should know that a simply velocity check is one the most rudimentary ones. Pattern matching systems, behavioural analytics and automatized analysis, all employed by Enbrite.ly are the new norm. These advanced techniques enable service providers and your company, working together, to achieve immediate results.
It’s important to note that coming up ways to inflict ad fraud is limited only by human imagination and it can appear in numerous forms. Here is a list of some known ways to cheat ad systems:
Velocity anomalies –too many impressions within a predefined timeframe – these are easy red flag events that are quite primitive and relatively easy to catch.
Example 1: Typical robotic traffic
Behaviour anomalies – strange mouse movements and touch movements – can be sophisticated, but still easy to catch.
Example 2: Typical robot behaviour from a previous measurement
Inventory anomalies – hidden ads, fake inventory created without content for the sake of showing ads only – can be caught by analysing traces of inventory manipulation and multiplication as well as checking dimensions.
Identity anomalies – ie. US targeted ad is being clicked in an Indian click farms – can be highly sophisticated if done right. However, by analysing micro traces, device fingerprints and uncovering the real source can solve these issues.
I’ve shared these thoughts as an attempt to raise awareness about the presence and importance of ad fraud and help you to ask the right questions when discussing strategies to tackle these issues.
We are eager to hear from you – give us a call, visit our offices, let’s have a discussion. We are always open for new perspectives.