Most sales teams review wins and losses by asking reps what happened. That is the least reliable way to find out. A win/loss analysis does one thing: it tells you what the buyer actually decided and why, not what the rep thinks happened.
Why rep self-report fails as a win/loss method
Ask a rep why they lost a deal. They will tell you: price was too high, timing was off, procurement took too long or the incumbent had a better relationship. These answers are not lies. They are the rep's interpretation of what the buyer communicated, filtered through what makes a clean story.
Ask the buyer and you may hear something different. The reason may not be price at all but the sales experience: a rep who was not prepared for their specific situation, a generic demo, or a technical question that mattered most to the person who made the decision and never got a good answer.
That is the gap. The rep's explanation and the buyer's can differ, and the buyer's version is the one that contains the fixable behavior.
When to run a win/loss analysis
Run a win/loss analysis whenever a deal above a size threshold closes, in either direction. Put the threshold where a conversation with the buyer on every deal above it is still practical; below it, the volume makes that hard. Ten to fifteen deals above it are a practical first batch for spotting systematic issues.
The worst time to start a win/loss program is after a string of losses. The data is useful then, but it is more useful when it includes wins. Without wins, you learn what not to do but not what to do instead.
Who to ask
Three groups give distinct perspectives.
The economic buyer. They made the final decision. Their account of when the decision was made, what tipped it and what the alternatives were is the most important data point in any win/loss review. They are also the hardest to reach, so make the effort anyway.
The champion. They advocated internally. They know which internal arguments worked, who pushed back and what concerns were raised that the rep never heard about. Champions are often willing to speak candidly after a deal closes in either direction.
The evaluator who said no. In lost deals, find the person who evaluated the product most closely and decided against it. Their objections are usually more specific and more useful than anything in the CRM notes.
What to ask
Avoid product-first questions. 'What did you think of our features?' produces feature feedback, not decision feedback. The questions below produce findings you can act on.
For won deals
Ask what moved the decision your way, and what nearly moved it elsewhere.
- When did you decide we were the right choice, and what made that moment happen?
- What concern did you have early in the evaluation that we addressed well?
- What almost made you choose differently, and why didn't it?
For lost deals
Of the three questions below, the last is the most useful in win/loss analysis. It asks the buyer to name the winner's behavior, not just their product. 'They had a better implementation plan ready on day one' is more actionable than 'their price was lower.'
- When did you start leaning toward a different option?
- What would we have needed to show or say at that point to change your view?
- What did the winner do in the sales process that we didn't?
How to turn findings into playbook changes
A win/loss analysis that ends in a slide deck changes nothing. The findings need to connect to specific rep behavior at specific call stages, and that connection needs to land in the team's playbook.
Four patterns to look for in your first ten to fifteen buyer conversations:
The discovery gap. Reps are not asking about the decision process early enough. Buyers describe making up their minds before the demo because the rep never gave them a reason to wait. Fix: add a decision-process question to the discovery playbook and train reps to ask it by minute 20.
The wrong proof point. Reps are using a case study that does not match the buyer's industry or deal size. The buyer went with whoever had a reference from someone like them. Fix: build proof points by industry and size into your sales materials, and in Sales Copilot attach the relevant one to the Prospect before each call.
The technical question that ended the deal. In enterprise deals, there is often a single technical question the rep could not answer well. The deal went to whoever answered it first and clearly. Fix: collect the technical questions that come up in lost-deal conversations and write approved answers for them into the playbook and the materials reps take into calls.
The late-stage stall. Buyers who chose a competitor describe a moment when the rep stopped advancing the deal and started waiting. Fix: add a playbook entry for what the rep does at day 15, 30 and 45 of a stalled deal.
How Sales Copilot supports win/loss analysis
A common weakness of win/loss programs is that they rely on rep recall. By the time a deal closes or is lost, the rep's memory of the call moments that mattered has faded. The exact phrasing is gone. The moment the buyer's tone shifted is gone.
After a live call, Sales Copilot by Cluely AI prepares a Debrief: a summary with recommended next actions, an analysis that points to evidence in the transcript, the transcript itself and a 0 to 100 readiness score beside its band. If the call saved an Account Action Plan, the Debrief shows it too. So the rep comes to a win/loss review with a record of what was said on each call, not a reconstruction from memory.
In practice, this turns win/loss analysis from a retrospective story into an evidence-based review. Ask specific questions, such as 'When the economic buyer raised the security concern, what did you say, and what happened next?' The answer does not depend on memory: the rep can open the transcript and check.
Read the Debriefs from several lost deals in the same industry side by side and you can compare what was said when the turning point came. That is more reliable than what anyone remembers in a deal review.
When the findings change what reps should say, put the change where the next live call can use it. In Sales Copilot, the Playbook and the Library materials you choose before a call shape the answers during it; in a Team workspace, share the updated Playbook and materials so everyone works from the same version.
For broader context on how the Playbook connects to live-call execution, see the AI sales copilot guide. For how to prepare the discovery calls whose Debriefs will feed the win/loss program, see the guide to discovery call preparation.
Frequently asked questions
What is a win/loss analysis in B2B sales?
A win/loss analysis is a structured program to understand why deals close or are lost, based on conversations with the actual buyers. It is distinct from pipeline reviews, which rely on rep self-report. In a win/loss program, the economic buyer, the champion and the key evaluator are asked about the deal after it closes. Their accounts of when they decided, what mattered most and what the rep did or didn't do are usually more useful than CRM notes or a manager's deal review.
How often should you run a win/loss analysis?
Run a structured win/loss review on deals above your size threshold, wins and losses alike, on a rolling basis. Ten to fifteen deals a quarter is a practical pace for spotting patterns. Do not review only losses: that gives a one-sided picture of what is broken, with nothing to compare it to. Reviewing wins alongside losses shows which rep behaviors are worth adding to the playbook and repeating.
Who should run win/loss conversations?
Not the rep who worked the deal. Buyers tend to be more candid with someone who was not trying to sell to them. Good choices are a sales manager, a revenue operations lead or a dedicated win/loss researcher. Some organizations use a third-party firm for enterprise deals where the buyer is unlikely to be fully candid with anyone from the vendor. The goal is honest feedback, not a relationship call.
What should you do with win/loss findings?
Connect them to specific playbook changes. A finding that says 'buyers want a reference call from their industry' becomes a playbook entry requiring the rep to line up an industry-matched reference before the demo. A finding that says 'reps were not prepared for the security question' becomes a set of approved answers added to the materials reps take into calls. Findings that do not connect to a specific rep behavior at a specific call stage change nothing. The output of a win/loss program is not a report. It is an updated playbook.
How does Sales Copilot help with win/loss analysis?
After a live call, Sales Copilot prepares a Debrief with recommended next actions, an analysis that points to evidence in the transcript, and any Account Action Plan the call saved. That gives the rep a record of what happened on each deal call instead of a memory of it. In a win/loss review, the rep can answer 'what happened when the security question came up' from the transcript rather than from reconstruction. Over time, comparing Debriefs from lost deals shows which moments repeat, and which Playbook and material changes are worth making.
Related guides
See how Sales Copilot supports the conversation. Start in your browser or download the app for Mac and Windows, and check pricing and our Privacy Policy before getting started.