Why interviews are not enough
Win/loss interviews are valuable and they do not scale. A good programme reaches a fraction of closed deals, weeks after the decision, through the people willing to talk. The rest of the story is already written down in systems you run.
- Coverage. Interviews cover a sample. Signals cover every deal.
- Lag. Interviews arrive after the quarter. Signals arrive during the deal.
- Politeness. Buyers soften the reason in an interview. On a call two months earlier, they said it plainly.
This article is a method for building the win/loss read from signals first, and using interviews to check it rather than to create it.
The sources that already hold the answer
- CRM notes and stage history. Where deals stalled, which competitor was logged, what the rep wrote the week before the loss.
- Call transcripts. Objections in the buyer’s words, competitors named unprompted, pricing pressure, the champion’s confidence.
- Support and onboarding tickets. For wins that churned early and for losses that trialled: what the product failed to do in the first weeks.
- Public reviews, yours and theirs. What buyers who chose the other tool say they got, and what they dislike about it now.
- Competitor moves. Pricing changes, launches and changelog entries dated against your losses. Many “we lost on price” stories start with a competitor move you never logged.
Building the loss taxonomy from evidence
Start with 40 closed deals, half won, half lost, across segments. Read everything the sources above hold on each, and write down the reason in the buyer’s words before you name a category.
- Group by the job the buyer was hiring for. Not by your feature list. “Needed to see churn risk per account” is a job; “Atlas” is a feature.
- Separate reason from trigger. The reason is why they chose; the trigger is what started the evaluation. A competitor move is often the trigger.
- Freeze a first taxonomy of eight to twelve reasons. Revisit it quarterly. If most losses land in “other”, the taxonomy is wrong, not the deals.
- Keep the quotes. Every reason carries at least one quote and the record it came from. This is what makes the read usable in a product meeting.
A weekly cadence instead of a quarterly report
- Every closed deal, tagged within the week. Reason, trigger, competitor, segment, ARR, and the quote.
- One view, by reason and by competitor. Wins and losses side by side, weighted by ARR, with the trend against last quarter.
- Moves on the same timeline. Date competitor moves against your losses to that competitor. If losses rise two to three weeks after a pricing change, you have the cause and the lag.
- Interviews to check, not to create. Interview the deals where the signals disagree with the rep’s reason. That is where the learning is.
Turning the read into battlecards that stay current
A battlecard is the loss taxonomy for one competitor, written for the next call. It should carry the buyer’s words, the competitor’s latest moves, and the counter, and it should change when the evidence changes.
- One card per competitor, one section per reason. The reasons you lose to them, in their buyers’ words, and what to say.
- Date every claim. A weakness from last year may be fixed. Prefer the last quarter.
- Hold for review. A drafted card is a draft. Product marketing approves before sales sees it.
Measuring the programme
- Coverage. Share of closed deals with a tagged reason and a quote. Target every deal, not a sample.
- Time from move to counter. Days between a competitor move and the updated battlecard. This is the number to cut.
- Win rate by reason. Whether the reasons you invest in move. If “needed X” losses fall after you ship X, the taxonomy earned its keep.
- Interview agreement. How often the interview confirms the signal read. Rising agreement means you can interview less.
Where Recon fits
The method above is a week of work to set up and an hour a week to run for a small team. It stops scaling when the number of deals, competitors and sources grows, and that is the case for an agent.
In HyperOrbit, Recon watches competitor sites, changelogs, pricing pages and reviews every day, scores each move, and drafts a battlecard for the critical ones with the evidence and a confidence score, held for review. Because it shares a signal layer with Chorus, a competitor move sits on the same timeline as the objections in your calls and the complaints in your tickets, so the lag between move and reaction is measured rather than assumed. In the worked example on this site a free-tier launch was followed 8 days later by 31 more SMB pricing complaints touching $214K of ARR; the methodology page says exactly what that figure measures. Enterprise plans include win/loss and battlecards for up to five competitors.