Home

How to use anomaly detection to seal revenue leakages due to price errors

Price errors and their impact:

Price errors are one of the ways in which revenue leakage occurs in e-commerce business. Although retailers put various checks and balances in place, pricing errors are still common. Data entry mistakes, misplaced decimal points, reversal of digits, and other clerical errors made in hurry are the major contributors. It can also occur because errors in feeding promotional offer dates. Promotions might unintendedly start early or end late. Such revenue leakages initially go unnoticed but gets detected after it becomes a significant issue.

In this digital economy, the cost of a price glitch is far more important than ever before. As soon as pricing error is spotted by someone, social media goes abuzz with the news, and it often goes viral. Once the post goes viral, the number of visits and orders exponentially increase. While the e-commerce business has the option to not honour such pricing errors, nowadays to save their online reputation, they end up honouring those mistakes.

Anomaly detection to the rescue:

Businesses must defend themselves against such errors as it eats into their already wafer-thin margins. Automated anomaly detection can help any dotcom monitor metrics such as sales price, sales volume, number of transactions, number of visits, and other elements in real time correlate them based on parameters like geography, demographics and behaviour. As soon as anomaly is detected at the KPI in a granular level, an alert reaches the right stakeholder. This insight provides the opportunity for appropriate action through which inadvertent losses could be significantly minimised.

‘First-class’ error:

In the recent past, Cathay Pacific had accidentally sold first-class airline tickets at economy prices between Vietnam and US. This news became viral in social media and resulted in significant number of people visiting the website and buying tickets. South china morning post tracked down eleven people who purchased 18 first class and business class tickets, which would have cost $685,800 at $27,000. It is unclear about the total number of people who have purchased the tickets. But they decided to honour the tickets as a new year gift. Two weeks down the line it happened again. This time tickets from Portugal to Hong Kong has a similar issue. £12,500 tickets were sold at £1,175. But this was spotted immediately. Their spokeswomen said very small number of customers have purchased the tickets and they are honouring it also. Using the power of anomaly detection, pricing errors can be detected in real time at a granular level and fixed before revenue leakage becomes catastrophic.

This article was originally published in CrunchMetrics.

Tags: