eSky and it’s Latin America’s brand eDestinos, covering globally almost 1000 carriers and over 1 million hotels, noticed that their rule-based fraud solution was not efficient in detection of fraud coming from the local attacks. The group reached out to Nethone for our Machine Learning models and dedicated Data Scientists. As a result, OTA benefited from higher approval rates and a significant drop in manual reviews. Also, the fraud management team easily achieved KPI despite numerous fraud attacks.
High chargeback ratios and multiple fraud attack incidents
The high-risk geographies that the group operates in generated attacks which had hefty costs. As is always the case, any set of static rules which might be used by a fraud prevention team has not been effective against attacks.
Instant denial and extensive manual reviews negatively impacted sales. Overall, the losses affected over 50% of all transactions.
Nethone team has developed a proprietary Machine Learning models tested to account for local singularities: differences in user's behaviour and payment methods. Data Scientists halved manual reviews while increasing the approval rate by 23% - boosting sales. eDestinos regularly talked with Nethone’s DS team to weight on insights and KPI alignment. It has resulted in the drop of the chargeback ratio by 6%. Overall the collaboration of Nethone with eDestinos increased the sales of the group.
Manual review rate