Product design question

If you are the Product Manager at Vistaprint in charge of Recommendations for the My Account section, how would you approach it?

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Product design for a specific surface: can you scope 'Recommendations' narrowly to My Account, define the user job, and design something measurable.

Clarify the surface: My Account (order history, reorders, saved items) rather than the homepage or search results.
Define the user job: a returning customer checking past orders or reordering business cards/marketing materials.
Identify the opportunity: recommend reorders (same product, refreshed design) and complementary items (e.g. business cards -> envelopes).
Propose the mechanism: recency/frequency-based reorder prompts plus category-adjacent cross-sell, shown on the account dashboard.
Address a risk: recommendations feeling irrelevant if purchase history is old or one-off (e.g. a single wedding order).
Define success: reorder rate and attach rate on recommended items, tracked against a control group.

Anchors on Vistaprint's actual business (small-business print/marketing products) rather than treating it like generic e-commerce.
Separates two recommendation types: reorder-the-same vs discover-something-adjacent, and prioritizes reorder first since intent is proven.
Flags the one-off purchase problem (e.g. wedding invites) and proposes filtering those out of the reorder algorithm.
Names a specific, testable metric: reorder conversion rate lift over a holdout group.

Designing a generic recommendation engine with no reference to Vistaprint's actual product mix.
Not distinguishing repeat-purchase customers from one-time buyers.

How would you avoid recommending irrelevant items to a one-time wedding-invite buyer?
How would you measure incrementality versus customers who would have reordered anyway?