Define the recommendation decision
A recommendation system chooses items for a particular user and context. Define eligibility before scoring: unavailable inventory or restricted content should not compete for a slot. Start with a popularity or recency baseline and specify whether success means useful discovery, completion, or another observable outcome. Clicks alone can reward attention without value.
Write an eligibility rule and a success measure for a course recommendation list.
Diagnose the stage that failed
Candidate generation narrows a large catalog; ranking orders the retrieved items; subsequent rules may enforce diversity or availability. A ranker cannot recover an item that never entered its candidate set. Measure retrieval coverage separately from ranking quality and inspect examples before replacing the entire system.
Of 100 relevant items, retrieval finds 70 and the final list contains 20. Explain why improving the ranker cannot retrieve the missing 30.
Treat missing interactions carefully
An item without a click is not necessarily disliked: the user may never have seen it. Exposure and position influence logged behavior. Track impressions where appropriate and evaluate using a time split that prevents future interactions entering user features. New users and new items need explicit fallback behavior.
Explain why treating every unseen course as a negative example can distort a learner preference model.
Evaluate the complete experience
For a user with four known relevant items, returning two of them in a five-item list gives precision at five of 2/5 and recall at five of 2/4. These metrics depend on the relevance labels and candidate universe. Add coverage and fresh-item slices, then test whether the online experience improves the intended outcome.
Calculate the two metrics, then explain what neither tells you about genuinely new items.
A relevant item never reaches the ranker. Which stage should you investigate first?
Keep exploring
Use the official certification page and its linked resources to validate your coverage. For guided labs, explore the Google Skills learning path. Revisit the practical exercise and explain your decision in your own words before moving on.