Measuring whether a single Zoom meeting succeeded: which signals to track

Zoom wants to know whether the meetings people hold on it are going well. Which data points would you track to decide whether an individual Zoom meeting was successful? Explain how you would define success for a meeting, which signals you would collect, which few you would rely on most, and where those signals could mislead you. ```hint Walk through one meeting Trace a meeting from the moment someone clicks the invite link to the moment the last person leaves. At each step, ask what could go wrong and what success would look like to the host and to the attendees. ``` ### Clarifying Questions - Who uses the answer: Zoom's product team judging meetings across the whole platform, or a host or company admin who wants a verdict on their own meetings? - Does "successful" cover whether the meeting achieved its purpose, such as reaching a decision or teaching a class, or only that it ran smoothly and people took part? - Which kinds of meetings are in scope: one-on-one calls, small team meetings, large all-hands, webinars or classes? Should personal and business use be judged the same way? - May we use meeting content such as transcripts, or only metadata and in-meeting events like joins, mutes and chat counts? ### What a Strong Answer Covers - A working definition of a successful meeting that separates getting in, audio and video quality, participation and outcome, and adjusts for meeting type and size. - Concrete data points for each part of that definition, with how each would be captured and normalized for meeting size and length. - A short, prioritized set of metrics (or a meeting-level score) with a reason for each choice and a way to check it against how participants actually felt. - The blind spots of the chosen metrics, privacy limits on what can be measured, and guardrails against misreading them. ### Follow-up Questions - Leadership wants a single "meeting success score" for every meeting. How would you build it and check that it matches what participants experienced? - Your participation metrics say most attendees of a company all-hands barely interacted, yet the host rated the meeting highly. How do you reconcile the two signals? - How would your metrics differ between a large webinar and a one-on-one call?