Skip to main content

Skills application solution

Worked example
In Progress

Skills application solution

Continue building your understanding with the content below.

Skills application solution illustration

Skills application solution​

Compare your skills application output to the solution example below provided by Multiverse subject matter experts.

** Solution**

** Step 1: Understand the context** The ML model predicts whether a user will complete a recommended course. The goal is to improve course completion rates and reduce unnecessary marketing spend by identifying users who benefit from targeted promotions.** Step 2: Define the error trade-off**

  • ** False positives**—promoting to users who would complete the course anyway—waste budget but have limited downside.

  • ** False negatives**—missing users who would benefit from a reminder—represent lost opportunity. However, marketing has a fixed budget, so limiting spend on uninterested users is more urgent.

  • Therefore,** false positives are more costly** , making** precision the higher priority** .** Step 3: Select your metrics**

  • ** Precision**: This measures how many of the users the model predicts as likely to need promotion truly benefit from it. High precision ensures that most users we promote to actually need the nudge, reducing wasted spend.

  • ** F1-score**: While recall also matters, F1 provides a balanced view that considers both precision and recall. It’s useful for assessing model improvements over time and understanding trade-offs.** Step 4: Communicate to stakeholders** We’re prioritizing ** precision** to focus marketing efforts on users most likely to need promotion. This ensures budget is used effectively and avoids oversaturating users with unnecessary messaging. We’re also tracking** F1-score** to monitor model balance and long-term performance improvements.

What this example does well​

  • Clearly connects model metrics to business priorities.
  • Justifies metric choices with practical trade-offs.
  • Uses concise, stakeholder-friendly language to explain technical decisions.

Tips for applying this skill in your role.​

  • Always start with the business outcome before choosing evaluation metrics.
  • Ask stakeholders which mistakes are more acceptable—and which are not.
  • Choose metrics you can easily explain and defend to non-technical audiences.
Callout icon

Action item: Reflection

Before moving on, take a few minutes to reflect on the following:

  • What did you do well?
  • Where could you improve?
  • If your chosen metrics were different, how would you explain your reasoning?