Case Study

Choosing a Quarterly Target from Years of Operational Experience

Case context: A company has maintained a medical transcription service for many years. The engineering team is now deciding what accuracy target to set for the next release cycle.

Question: According to the underlying idea, how should the team choose the target, and what kind of judgment should guide it?

Sample answer: The team should use its long history with this system to estimate what improvement is realistically attainable in the next cycle. Because they have seen the system’s behavior over many releases, they can make a practical judgment about likely progress, and that judgment should be used to choose the target accuracy.

Key points:

  • Years of experience with one application can support realistic forecasting
  • The team should estimate how much improvement is feasible in the coming cycle
  • That estimate should shape the target they set
  • This is about operational performance, not about user-facing product taste

Rubric: Full credit for explaining that long-term experience with the system should be used to judge realistic near-term improvement, and that this judgment should determine the target accuracy.

0

1

Updated 2026-08-12

Contributors are:

Who are from:

Tags

Machine Learning

Deep Learning

Supervised Learning

Dive into Deep Learning @ D2L

Data Science

Machine Learning Strategy

Machine Learning Yearning @ DeepLearning.AI