News as it is -

Best Media Info

Partner Content

How Flipkart deals with ad fraud and brand safety

At MMA’s road show series that kicked off in Delhi on March 13, Rahul Pandey of Flipkart explains how the company deals with ad fraud and ensures brand safety

Ad fraud is a major concern in the industry and fake installs of apps, clicks and views have led to inaccurate measurement of digital presence, warned Rahul Pandey, Digital Marketing Director, Flipkart.

Pandey, the keynote speaker at the Mobile Marketing Association’s (MMA) first-ever road show in India that began in Delhi on March 13, pointed out various facets of ad fraud and a few ways to ensure brand safety. He explained how Flipkart deals with ad fraud and brand safety.

Pandey said the rate of mobile ad fraud in India is about 31.9%, as compared to the world’s average of 16%. Globally, ad fraud was about $19 billion in 2018, and is expected to grow to $44 billion by 2022.

In India, majority of the ad fraud takes place in the digital e-commerce sector, which is about $0.8 billion (48.2%), followed by the leisure and travel sector, which is around $0.4 billion (26.3%) and entertainment and gaming, which is roughly about $0.2 billion (14.6%).

Talking about the evolution of ad fraud over the years, Pandey said basic installs (apps), clicks and views have been faked, which lead to inaccurate measurement of digital presence. Also performance events are faked, which culminates into fraudulent sales, users, registrations, etc. The entire digital behaviour is also being faked, which affect re-engagements, brand safety, equity, relevancy and reputation.

He pointed out that ad fraud takes place across channels -- SEM (search engine advertising), MM (mobile marketing), affiliate marketing and RT (re-marketing).

As per one of the case studies presented by Pandey on mobile marketing, the topmost and smallest portion of the pyramid comprises active users, and getting down on the lower and a little wider portion is the set of lapsed users, and finally the bottommost and largest portion of the pyramid consists of the inactive users. Only small percentage of installers is converting into customers.

The frauds include device fraud (35%), click injection (30%), incent fraud (16%), CTIT (14%) and IP fraud (5%).

The key interventions for mobile marketing that were utilised were -- optimisation based on sub-publisher ID, working on cost per action model, use of fraud tool to identify/prevent fraud and building custom post back rules to prevent fraud.

After the interventions, the cost per installation reduced, and the average revenue per installation increased.

The second case study presented by Pandey was on that of affiliate marketing, wherein he said the content went through three stages -- from Flipkart to the internal and external tools and then to the publisher, thus keeping the cycle between the Flipkart and the publisher running.

The types of fraud in affiliate marketing include attribution (44.92%), re-seller (20.61%), policy violation (12.97%), self order (10.82%), brand bidding (7.89%), banned affiliate (2.61%), emailer violation (0.15%) and internal affiliate (0.02%).

The key interventions in preventing ad fraud in affiliate marketing were -- building tech solutions for right attribution, algorithmic detection of re-sellers based on patterns match, implementation of strict policies on brand safety and on-boarding of external tool for pro-active defence.

The impact on the key metrics of affiliate marketing was that there was a 12% saving on the monthly budget, an additional GMV to the channel and an improvement in key health metrics (RPC, UPC and LTC).

Pandey said some of the strategies that worked for Flipkart were a comprehensive cross-channel analysis. This included app and web fraud, a sophisticated approach wherein there was a focus on preventing performance faking event frauds, cannibalisation of organic or search, building internal and external expertise to proactively handle ad frauds on a large scale, and the continuity in not treating ad fraud as a one-time affair as fraudsters will keep exploiting every loop they can find.

Post a Comment