What-If Tool

This page presents a static what-if analysis based on the RPL model. It visualizes hypothetically how the improvement in each effort driver can impact RPL scores and ultimately translate into GBV gains.

  • Effort drivers in 3 buckets: Customer Time, Customer Steps, and Customer Hand-Offs.
  • Service Improvement: the hypothetical improvement of effort drivers, either in percentage or in absolute count. E.g., decreasing average queue time by 2%; reducing chat count by 1 across journeys.
  • Repeat Purchase Conversion Rate: percentage of customers converting from non-repeat to repeat , based on 1.58 million randomly sampled journeys from 2021. This is the output by the model.
  • Applicable Journeys in 2021:

    • Only include customers who have less than 156 3-year order counts, to exclude potential travel agents.
    • Journeys with a positive value in the effort driver of interest. E.g., for queue time analysis, applicable journeys must have a total queue time > 0 so that it has room for improvement.
    • Percentage of Applicable Journeys in 2021 = Number of Applicable Journeys/Total Journeys. E.g., 46.83% of journeys in 2021 had a queue time greater than 0.
  • Historical Order and GBV Data

    • 3-Year Order Count and 3-Year GBV: based on average values of customers who initiated journeys in 2021.
    • Average dollar value per order is obtained by dividing 3-year GBV by 3-year order count.
  • Expected GVB Gain

    • 3 Months GBV Gain = Repeat Purchase Conversion Rate * Number of Applicable Journeys * Average Dollar Value per Order
    • For instance, if there's a 1% decrease in total queue time, 0.04% of 7,502,476 journeys will convert from not having a repeat purchase to having a repeat purchase, which is ~3000 new repeat orders. The average GBV value of 1 order is ~$470, 3000 orders will result in approximately $1,412,231 in GBV gain.
    • 1 Year GBV Gain = (Repeat Purchase Conversion Rate * Number of Applicable Journeys * 3-Year GBV)/3
    • 3 Year GBV Gain = 1 Year GBV Gain * 3 RPL WhatIf CustomerTime RPL WhatIf CustomerSteps RPL WhatIf CustomerHandOffs
On this page