Controlling Advertising Wastage: Bigger Side Effects of Bots

Bigger Side-effects of Bots

This article written by me appeared in Brand Equity on 22-11-2017. 

Bigger side-effects of bots

Bot traffic damages go beyond just media investment wastage

Digital ecosystem is plagued with digital ad fraud and it is not slowing down. The two major types of fraud are Domain Spoofing and Bot Traffic. Within these two alone there are so far up to 60 types of different implementations that are among the known ones. Domain Spoofing is unauthorized use of a third-party domain name in an e-mail message in order to pretend to be someone else. And bots, in this context, refers to software applications run by malicious organisations that are designed to increase the page views/engagement of an advertising campaign. Why malicious? Because while the client assumes he is paying for human prospects, in reality these are machine led applications behaving like humans. Multiple studies from IAS (Integral Ad Science), Tune Mobile, DoubleVerify and the likes, indicate that the extent of wastage from bots can be as high as 30%. Some of the readers may know that this investment has a zero ROI but did you know it can even lead to the ROI being negative?

The implication of driving a bot initiated click does not end with just the media cost of the click but extends beyond that. Distorting your website analytics data which gets skewed owing to the bots traffic. That’s right. Your interpretation of the audience insights such as their geography, age and interests are now all skewed. A wrong interpretation from site analytics leads to wrong audience identification which further leads advertisers to target the wrong audience. We observed this in a recent project with one of our clients, where we realised that the client was targeting the wrong audience as there was a drastic change in the demographic breakup after we blocked the bot traffic from hitting the client’s website. The revised plan not only improved the planning efficiency by 15% as a result of lower eCPC (effective Cost per Click)  but also demonstrated a 14% lower bounce rate. And this does not even account for the direct savings of 18% due to reduced bandwidth and server load. The bot analyser, powered by Artificial Intelligence, is now a resident tool on the client’s website which continuously updates itself with globally discovered bots on a real-time self-learning mode.

However, as the tracking mechanisms begin to get smarter, so do the bots. For example the more sophisticated bots can imitate humans to scroll on the website or even do a spam form fill.  At first, it might sound intimidating to tackle such a huge technological problem which is eating into media money, but with efficient tracking mechanisms, the problem can be mitigated to an acceptable level. Following tips can be used by advertisers to counter bot traffic:

• Accredited tools can be employed to ensure media being  bought only on transparent platforms with pre-set KPIs and appropriate clauses for payment, basis agreed KPIs.
• Direct bot traffic can be controlled by deploying effective bot detection and protection techniques which provide for a cleaner analytics data for more effective insights. 
• Being overly cautious while investing in programmatic platforms which have comparatively higher instance of ad frauds than direct buys across advertising formats. 
• Implementing transparent programmatic planning and buying model where all details are available for inspection.(Often agencies and DSPs/SSPs (Demand/Supply Side Platforms) will make clients believe it’s a black box, but it isn’t)
• Building  and managing a whitelist (trusted sites)
• Since most humans sleep during the night J, bot percentages are high between 11PM-5AM. Such hours, to be avoided.

Now taking a leap from following the media benchmarks to questioning and dictating those benchmarks. While one can’t completely get away from bots, existing solutions can reduce wastage by up to 85%,as per a recent IAS study . That is encouraging enough to start the practice, I would reckon.