Buying ad impressions in real-time is a rapidly growing sector in the advertising industry. DSPs, SSPs and Exchanges all communicate via RTB (real-time buying) to buy and sell in auctions on an impression-level. Data providers enhance the information of an impression with specific user data. Efficiency, transparency and scale are the predominant advantages of this development. But there are still a few pitfalls when buying programmatically.
This might seem strange considering I have just said that transparency is one of the predominant advantages of programmatic buying. It’s true, you are able to see the website before bidding for an ad impression, but as there are millions of websites there are unfortunately some black sheep who generate bot-traffic to tap into the auctions. According to Accordant Media this kind of fraud is seen on 10.3 % of the whole programmatic inventory.
But advertisers aren’t the only fraud victims. Publishers can get attacked by bots in a high-quality environment (i.e. video) that then fake an auction. They want to collect the user’s cookies, which they can reuse in other bot-traffic sites.
There are also other technical methods like Pixel stuffing (an ad is shown on a single pixel) or ad stacking (several ads are put on top of each other). In the end the total fraud volume in programmatic buying can be anywhere up to 22%.
With good filter techniques, blacklisting and monitoring most fraud can be prevented. A good ad operations team and a proper platform can minimize the amount of fraud by 85%.
2. User Frequency
RTB is still mainly used for Retargeting. In order to move the user to a conversion, they are shown an ad featuring the product they were recently interested in. This can mean a user sees an ad more than 500 times, resulting in ad fatigue.
One reason is that many advertisers still test and compare DSPs, meaning that the end user sees ads from both systems. The targeting is redundant and one platform does not know what the other is doing or has already done.
Retargeting still performs very well, but due to this setup and the resulting ad fatigue, the margin gets lowered.
To understand and better control this phenomenon advertisers should define and control KPIs that take into account the frequency of a user group (i.e. heavy index user and low index user). Also, the use of dynamic ad creatives and the change of creative cycles should also be used to lower ad fatigue.
3. Reach redundancy
In the end, I am my biggest competitor. Assuming that I am using several DSPs or Ad Exchanges or Ad networks (directly or indirectly), this can mean that all platforms are competing for the same impression. Even a second price auction isn’t helpful anymore, I will pay my maximum bid and potentially decrease my margin.
The problem is similar to the user frequency problem. RTB is an open standard and people are talking about an efficient market like the stock exchange. Paradoxically, RTB actually works best when one uses as few systems as possible in order to minimise redundancy and understand anomalies better.
Does that mean it’s already time to consolidate these systems?
Feature image: Man holding credit card in hand and entering security code using laptop keyboard via Shutterstock. Copyright: Kostenko Maxim