Case Interview: Measure Success
Case Interview: Measure Success
1. Analytics Framework
For metrics selection, the funnel anaysis model is the most frequently applied type of frameworks. In this case, we commonly applied the AARRR framework to allocate proper metrics. For details on models and frameworks, refer to ongoing
The according metrics of each objectives under various business models are listed below:
Business Model | Acquisition | Activation | retention | Revenue | referral |
---|---|---|---|---|---|
E-commerce | |||||
SaaS | |||||
Free Mobile APP | |||||
Media Site | |||||
UGC | |||||
TS Marketplace | |||||
For all | DAU/MAU, DNU by platform |
2. Criterion for Metrcis Selection
3. Procedure of Measure Success Problem
This type of cases usually come as "find a metric to measure the extent.....(how good a new feature/strategy/product is)". IT evaluates one's ability to design a measure according to certain business objectives
Clarify the objective:
Ask questions about:
- How this feature works?
- who is it for
Confirm the business objective
according to the explanation from the interviewer, conclude the objective of the feature to the following aspects:
- Acquisition: bring in more user, help them discover the product
- Activation: make user taking actions we want them to, keep users engaged
- conversion and retention: keep user continuing to use the product, convert them from free-user to paid user
- Money: make revenue from users
Select metrics according to objective
Objective Acquisition DAU/MAU, DNU by platform, Activation ACU/PCU, PV/UV, time spending, interaction behaviour Conversion and Retention CTR,CO, pay rate, 7 day retention rate/churn rate Money ROI,GMV, AOV, ACV, ARPU/ARPPU Select 1 main success metric, this metric should be directly relevant to the business objective
Select 1 second success metrics, this metrics should be a supplement. It improve the completeness of the logic relationship of the key metrics and business goal. To fing the second goal, consider the following questions:
- Can the change of key metrics be caused by factors other than the feature?
- Is the key metric equivalent to the business goal? Is there any intermediate process
- If the key metrics changes on a group level, should we have same changes on single user level?
Alternatively, we can do a funnel analysis to find the secondary success metrics
Select 1 guardrail metric, this metric is critical to the main goal of the whole product. The feature should not have significant negative effect on it
structure the answer
if the guardrail metric shift towards a negative direction, the feature is probably not success.
Otherwise, if the key success metric and the second both shift towards a positive direction, the feature is successful.