Key takeaways
- Sample won, lost, no-decision, disqualified, and stalled opportunities across segments and deal conditions.
- Use an interviewer who was not responsible for the deal and interview close enough to the decision for the sequence to remain recoverable.
- Reconcile buyer accounts with CRM, proposal, activity, and delivery evidence without treating any one source as truth.
- Report coverage, nonresponse, contradictions, and limits beside every recommendation.
01
“We lost on price” is often the end of a CRM dropdown, not the explanation
The salesperson records price after a difficult loss. The buyer remembers unclear implementation risk. Procurement remembers an incomplete security answer. The executive sponsor remembers that the internal project lost priority. All four can be true, and a single loss-reason field cannot preserve the sequence.
Define the population and draw a sample before asking account teams whom to contact. Include no-decision and stalled work; those outcomes expose friction that competitor-focused analysis misses. Stratify by market, service, deal size, sales motion, tenure, and outcome only where the sample can support meaningful contrasts.
Swipe to compare every column
| Source | Useful evidence | Built-in blind spot |
|---|---|---|
| Buyer interview | Decision sequence and perceived risk | Recall and diplomacy |
| CRM | Timing, stages, fields and ownership | Seller-entered categories |
| Proposal and security files | What was actually offered | Internal discussion missing |
| Delivery record | Whether the promise held | Available only after a win |
02
Protect independence without turning research hostile
Use an interviewer who did not sell the deal. Explain that the conversation is research, not a recovery call. Do not reveal an account manager’s hypothesis before the buyer describes what happened. Ask for trigger, shortlist, evaluation, internal participants, turning points, concerns, final decision, and what nearly changed it.
Invite account-team context after the independent account is captured. Differences are findings. They can reveal missing stakeholder access, assumptions that never reached the buyer, or buyer rationalization after the fact.
03
Make nonresponse part of the result
Track invited, reached, declined, completed, and unusable interviews by cohort. AAPOR treats nonresponse and lifecycle quality as core survey concerns. Win/loss interviews are not a probability survey, but hidden recruitment failure still distorts what the team hears.
Do not force saturation across a heterogeneous market. Report where evidence is dense, thin, or absent. One vivid loss can justify investigating a failure; it cannot establish its frequency.
04
Turn themes into owned tests
Write each finding with supporting cases, countercases, likely mechanism, affected cohort, confidence, and unresolved question. Link it to a small action: change security sequencing, publish implementation evidence, clarify scope, involve a user earlier, or retire a weak qualification rule.
Review the action after a defined set of later opportunities. The purpose is not to grade sales or produce a quarterly slide of top reasons. It is to change one part of the buying experience and see whether the relevant friction becomes less common.
Primary sources and further reading
Use the source material to validate details against your own context and current platform configuration.
- AAPOR: Best Practices for Survey Research
- Malterud et al.: Sample size in qualitative interview studies
- NIST: AI Risk Management Framework—Map and Measure
This field note follows the XenGrowth editorial policy: primary sources where available, visible limitations, material review dates, and no invented first-hand experience.
Stay with the problem
