Uber Eats experiment: does nudging customers toward healthier orders raise engagement?

You are the product manager for Uber Eats. Take as a working assumption that helping customers eat more healthily would make them more engaged with the app. Propose a new feature you would build to find out whether that assumption is true, and explain how you would use it to test the assumption. ```hint Two links in the chain The assumption only holds if the feature first changes what people order, and the healthier ordering then changes engagement. Make sure you can observe each link separately. ``` ### Clarifying Questions - How should engagement be measured for this test: order frequency, retention, time in the app, or spend per customer? - Should the feature work through the product experience (ranking, labels, tools), or are price levers such as discounts on healthy dishes allowed? - Does Uber Eats have reliable nutrition data for menu items today, and in which markets? - Is the target every Uber Eats customer, or a group such as people who already order healthy food or have said they want to? ### What a Strong Answer Covers - The assumption restated as a testable hypothesis, with a defined engagement metric and a definition of healthier ordering. - Customer segments and the obstacles to eating healthily through food delivery, with a chosen focus. - Feature options compared on how cleanly and cheaply they test the hypothesis, ending in one clear choice. - An experiment design: randomization unit, primary and guardrail metrics, duration, and decision rules for confirming or rejecting the assumption. - Risks such as inaccurate nutrition labels, effects on restaurant partners and a tone that feels judgmental. ### Follow-up Questions - Customers in the test ordered healthier meals, but their order frequency did not change. What do you conclude, and what do you do next? - How would you handle restaurants that publish no nutrition information? - Engagement rose, but average order value fell. Would you launch the feature?