Launch retros that actually change the next quarter

Most launch retros devolve into praise sandwiches. We timebox three lenses: inputs (briefs, research, design), throughput (approvals, tooling, meetings), and customer signals (support spikes, sales objections, analyst questions). Each lens gets a facilitator who cannot defend their own function—rotation keeps blame from clustering.

Capture decisions as experiments. If the retro concludes that approvals were late, assign one measurable change: for example, a single approver for tier-2 copy or a hard SLA for legal review hours. Revisit the metric next retro; if it did not move, the experiment failed and you try a different lever.

Share a one-page retro brief externally when appropriate. Prospective students in product marketing fundamentals care that you treat process seriously, not that every launch is flawless.