Respond to customers using a tenth of their reserved database capacity

Suppose you are the product manager for a cloud database offering at Google. Customers are allotted a number of capacity units called slots. A usage review shows that your customers are, in aggregate, using only about 10 percent of the slots allotted to them. What would you do? ```hint Question the number first Decide what the 10 percent actually measures before proposing any fix: an average or a peak, all customers or a handful, and over what window. The right response depends on what sits behind the figure. ``` ### Clarifying Questions - What exactly is a slot, and do customers pay for the full allotment whether or not they use it, or only for what they consume? - Is the 10 percent an average across everyone, or do a few heavy users hide a long tail of nearly idle accounts? - Is low utilization new, or has it always been this way, and did a pricing change, release or sales push happen around the time it appeared? - Is the goal customer value and retention, higher revenue, or freeing capacity that sits idle? ### What a Strong Answer Covers - A precise definition of utilization and a check that the measurement itself is correct. - A structured set of hypotheses split between customer-side causes and Google-side causes, and a way to test them with data. - Customer segmentation to show where the problem is concentrated. - Concrete remedies matched to the confirmed cause, with their revenue and customer trade-offs. - Metrics that show whether the fix works, with guardrails. ### Follow-up Questions - Data shows that most customers are over-provisioned by roughly the same amount at renewal. What would you change in pricing or packaging? - A recommendation to right-size would lower committed revenue this quarter. How do you make the case to leadership? - How would you test your remedy before rolling it out to all customers?