Revenue / Atlas models achieved 40% revenue contributions

[bstiawan]

Description

• Atlas' new model has been deployed for almost a month and has started executing price prediction starting from this month.
• There is a need to quantitatively and objectively measure the performance of Atlas contribution towards revenue as a whole.

Problem

1. ATLAS does not have clear contribution metrics that can be used to measure its current performance and target performance upon creating a milestone.
2. ATLAS does not have a clear SLA to be achieved after it's deployment after several months.

Solutions

• The measurement of ATLAS revenue contribution should be based on the achieved revenue under fair comparable circumstances. That is, the achieved in-month revenue, i.e. from in-month pricing decisions. issue
• The measurement of ATLAS and revenue management as a whole should be measured as the percentage contribution of revenue from the in month potential revenue. issue

Measurement metrics

Dashboard Link

• Atlas revenue contribution from in-month generated revenue Query Link
• (Optional) Revenue management revenue contribution from in-month potential revenue

Current measurements:

• No current measurements since there were no metrics.

SLA

• [x] Atlas revenue contribution from in-month generated revenue: 40%
  • Atlas consisting of Random Forest and Embedding DNN achieves 79.2% of in-month revenue.

• [ ] (Optional) Revenue management revenue contribution from in-month potential revenue: 80%
  • [ ] Currently revenue management effort contributes to 64.23% of in-month potential revenue

Updates:

• 2024-04-19: Consolidated the KPI and SLA for Atlas and Revenue management with Felix, Widya, Bella

[Vidiskiu]

Overall Point: 4.4

Functional Complexity: 1

Developing clear metrics for assessing revenue contributions requires understanding of both the business context and the data involved, with moderate functional complexity.

Technical Complexity: 0.7

The task involves analyzing existing data and implementing new measurement methods, which is technically complex, but does not involve building new systems from scratch.

UI/UX Complexity: 0.5

Some UI elements might be necessary for displaying the revenue contributions metrics on dashboards but it's not the core of the issue.

Data Manipulation: 1

There will be significant data manipulation needed to accurately measure revenue impacts, including possibly creating new queries and handling data from various sources.

Testing: 0.1

While important, testing in this context largely involves validating the accuracy of data and revenue predictions, not extensive QA processes.

Dependencies: 0.1

The issue does not appear to require additional dependencies outside the scope of existing data management systems.

Risk and Uncertainty: 0.1

Risk is minimal as the issue pertains to the measurement of existing models' performance instead of implementing new, untested features.

User Impact: 0.9

Having accurate metrics for revenue contribution has a high impact on business users who rely on this data to make informed decisions, but it does not directly affect end consumers.