Prioritizing ten Google Cloud customer customization requests when only one can be built

You are a product manager for a Google Cloud product. Ten customers have each asked your team to customize the product to fit their needs, but your engineering team has capacity to take on only one of these requests. Walk through how you would evaluate the ten customers and their requests, and how you would decide which one to build. ```hint Look past the requester For each request, ask who else would benefit if it were built, and what it would cost the company to say no to the customer who asked. ``` ### Clarifying Questions - Does the capacity for one request cover this quarter or a longer period, and does it compete with work already on the roadmap? - What do we know about each of the ten accounts: current spend, expansion potential, renewal timing, and whether any is at risk of moving its workloads to another cloud? - Do any of the ten requests overlap, or ask for something the product could offer every customer as a configurable feature? - Is extra delivery capacity an option, such as professional services, a partner, or the customer funding the work? ### What a Strong Answer Covers - A stated objective for the choice that ties to the product's strategy, rather than defaulting to the largest account. - Understanding the need behind each request, and filtering out requests that existing features, other Google Cloud services or the current roadmap already address. - Explicit evaluation criteria (reach beyond the requester, account value and risk, strategic fit, effort and maintenance cost) applied consistently to all ten. - A stakeholder plan: aligning sales and customer success on the criteria, and handling the nine customers who are not chosen. - How the success of the chosen build is measured. ### Follow-up Questions - The highest-spending customer says it will move workloads to a competing cloud unless its one-off request is built. What do you do? - Three smaller customers ask for variations of the same capability. How do you compare that cluster with one large account's request? - After launch, how would you judge whether you picked the right request? - Suppose your team can help three of the ten customers instead of building one request. How do you choose a set of three that works well together, and what is your plan for the seven customers you turn down? - Suppose the prompt only says your team can "help" these customers, without saying how. What would you clarify about what help means (a new feature, removing a product limitation that blocks them, or hands-on engineering or migration support), and how would each answer change your approach?