A lost-deal review should produce a decision about what to test, not a story that makes the loss feel inevitable. The useful unit is a dated sequence of buyer statements, seller actions and unresolved questions.
Define the cohort before explaining the losses
Define the period, segment and loss stage before looking for patterns. Include no-decision outcomes separately from competitor selections. Keep open deals out of the lost cohort and flag records missing a close date or reason rather than silently excluding them.
Reconstruct evidence without filling in the unknowns
Collect the proposal version, buyer feedback, meeting notes and approval requests. Distinguish direct buyer statements from a representative’s interpretation. Record when an item was requested, when an internal team received it and when the buyer received a response. Respect recording permissions and access limits.
Use categories such as product gap, budget, timing, internal delay and unknown. Multiple causes may apply. Ask what evidence would contradict the proposed explanation. A slow response before a loss is an association, not proof that faster follow-up would have won the deal.
Hypothetical example: Suppose a fictional team reviews ten losses: four buyers explicitly cite a missing export, two mention timing, and four have no confirmed reason. One of the export cases also waited six days for an answer. The manager sends the product gap to product planning and tests response routing separately. The team does not relabel the four unknowns as price objections to make the chart complete.
Choose one change the team can actually test
Choose one change, owner, cohort and review date. For example, test a technical-question intake template and measure time to a verified answer. Record other changes in pricing, lead quality and product availability. Share conclusions with uncertainty rather than turning a small review into a performance ranking.
- Cohort and loss definitions fixed.
- Buyer evidence distinguished from interpretation.
- Unknown and multiple causes retained.
- One experiment has an owner and measure.
- Results discussed without causal or revenue guarantees.
What a lost-deal review cannot establish
Buyers may decline an interview or offer incomplete explanations. Small cohorts and selective CRM notes limit comparisons. Avoid blaming individual sellers based on missing evidence, and do not contact buyers repeatedly after they decline participation.