Diagnose What Drags Down Uber Driver Ratings in One Area and Lift the 4.5 Average

You are a product manager at Uber. In the area where you live, the average rating that riders give their drivers currently sits at 4.5 out of 5 stars. How would you raise that average? ```hint Look at the distribution, not the mean An average on a 1-to-5 scale that already sits near the top can move for very different reasons. Work out which slice of trips is pulling it down before you propose a fix. ``` ### Clarifying Questions - Is the 4.5 an average over every rated trip in the area, or an average of each driver's own rating, and over what time window is it measured? - How does 4.5 compare with other areas of the same city: is this area an outlier, or somewhere in the middle? - Do we already know anything about why riders here leave low ratings, such as rating reason tags, written comments or support tickets? - What target and time frame are we working toward, and are any levers off the table, such as removing drivers from the platform? ### What a Strong Answer Covers - How the average is built, and why a few one-star trips can outweigh many four-star trips. - A structured diagnosis that segments low ratings by driver, trip type, time, place and stage of the trip before choosing a fix. - Separating causes the driver controls from causes the app or the city creates, and routing each to the right owner. - Prioritized interventions for drivers, riders and the product, with a way to test them. - A success metric plus guardrails that catch rating inflation or damage to driver supply. ### Follow-up Questions - Your changes lift the average to 4.6, but trips per rider in the area fall. What might have happened, and what do you do next? - Many low ratings turn out to follow pickups where the map pin was in the wrong place. Should those ratings count against the driver? - How would you test a driver-side program when drivers and riders in the same area affect each other's experience?