The company runs a large online auction platform and shopping website that connects buyers and sellers in over 190 countries. Their website has millions of sellers with 1.9 billion global listings and over 132 million active buyers.
The company’s online auction platform and shopping website allows customers to browse through different categories of products or search results, click on individual listings that they are interested in and then purchase the required items. The website has a set of algorithms and models which determine the order handling time for each customer order. This order delivery time is dependent on the customer’s location, It also takes into consideration pre-defined lane rules which are applied to the complex models.
Accordingly, during a customer transaction, the website displays the estimated shipping time and delivery date for each product on the
– Item page
– Checkout page
– Successful Transaction page
Different sets of business users and Subject Matter Experts (SMEs) noticed that there were occasional mismatches between the displayed delivery dates on the Item page, the Checkout page, and the Successful Transaction page. The business owners did not know if this was being caused by any contradicting rules in their order handling algorithms and models. Additionally, they were not aware of the frequency with which this problem occurred.
Several business users felt that the inconsistency in delivery dates could potentially result in significant lost revenues arising from
- Orders not being placed (if the Item page showed a later delivery date than the date which would eventually be shown on the Successful Transaction page) or
- Orders getting cancelled (Item page showed an earlier delivery date than the date which would eventually be shown on the Successful Transaction page)
The team of business analysts from Prescience Decision Solutions was tasked with understanding the scale of the date discrepancy issue between the Item page, the Checkout page, and the Successful Transaction page, for one of the company’s largest markets. Over a 2 month time frame, the team selected various product listings from sellers in that test market and simulated orders to different parts of the country and other countries, as well. For each iteration of their analysis, the team kept the logistics method the same, so that they eliminated any external variables which could influence the generated delivery dates.
During the engagement, the team found that,
– Between 5 – 6% of all simulated cart orders had a mismatch in the displayed delivery dates of the Item page and the Checkout page.
– Between 2 – 3% of all simulated cart orders had a mismatch in the displayed delivery dates of the Checkout page and the Successful Transaction page.
For all the above simulated orders with date discrepancies, the nature of the mismatches remained constant. Our team found that
– The delivery date on the Item page was higher than the delivery date on the Checkout page.
– The delivery date on the Checkout page was lower than the delivery date on the Order Successful page.
The findings from our team’s study were submitted to the various business users and SMEs. The scale of the mismatches in the published delivery dates and the consistent pattern which was recorded, supported the business user hypothesis that the company could be losing revenue on account of this issue.
The final report submitted by our team of business analysts helped the company’s business users and SMEs understand exactly how many customers were being impacted by the inconsistency of the displayed delivery dates. With the findings of the report, the company’s engineering team can narrow down the underlying issue which typically arises from a software bug or a mismatch in the business logic used while developing the order handling algorithms and models.