How to Run Lost Deal Analysis From Your Sales Calls (2027)

Your CRM says most of last quarter's closed-lost deals were lost on 'pricing.' The VP of Sales repeats it at the QBR, product shrugs, and nothing changes. At BuildBetter, we hear this from sales leaders, RevOps teams and founders all the time. They don't trust their closed-lost reasons, but they can't prove what actually went wrong. The cause is predictable. A rep picks the loss reason from a dropdown when the deal closes, often weeks after the real problem first appeared. 'Pricing' is the least embarrassing option. It ends the conversation and needs no evidence.

A better dropdown won't fix this. The evidence is already in your recorded sales calls, where buyers said what was missing, which competitor they were evaluating and what they needed before they could sign. This guide covers a lost deal analysis method that a team with a spreadsheet and no budget can run this week. It includes a worked example with real numbers and a clear point at which the manual version stops working.

What Is Lost Deal Analysis? (Definition)

Lost deal analysis is the practice of reviewing closed-lost opportunities, using what buyers said on sales calls rather than the reason a rep selected in the CRM, to find the real causes of losses, group them and measure how much pipeline each cause cost.

It is related to win-loss analysis, but the two are different. A traditional win/loss program compares won and lost deals and usually relies on post-decision buyer interviews or surveys, often run by a third party. Lost deal analysis from calls mines conversations you have already recorded. It is cheaper and faster. The tradeoff is that it only sees what was said on calls.

The method depends on one distinction:

• Stated loss reason: what the CRM dropdown or the buyer's final email says. Example: 'budget.'
• Observed loss reason: the blocker the call evidence supports. Example: 'needed SSO, raised it on three calls, never got an answer.'

When the stated and observed reasons disagree, the observed reason is usually the one you can act on.

A finished closed-lost analysis produces three outputs:

1. A ranked list of loss reasons weighted by pipeline dollars, not deal count.
2. A set of product gaps backed by dated, verbatim buyer quotes that product can take into roadmap planning.
3. A baseline built on a fixed loss reason taxonomy that you re-check next quarter to see whether your fixes worked.

Why Closed-Lost Reasons Are Usually Wrong

Closed-lost reasons in the CRM are usually wrong because they are captured once, from memory, in a single field, at the moment the rep has the least incentive to be precise. HubSpot ships a default 'Closed lost reason' deal property, and Salesforce teams usually add a custom picklist on the Opportunity object. In both, the field is normally limited to one value chosen when the deal moves to Closed Lost.

Five reasons CRM loss reasons mislead:

1. Dropdown bias. The list has six to eight options, and 'pricing' or 'no budget' absorbs everything ambiguous. One dropdown value can hide several unrelated problems.
2. Recency and memory. The reason is logged at close, but the deciding moment often happened on call two or three. Nobody goes back and re-listens.
3. Buyers soften the truth. 'It's a budget thing' is a polite exit. The actual blocker, such as a security requirement, a missing integration or a competitor's existing contract, usually came up earlier and in more specific terms. Gartner research finds that buyers spend only about 17% of their buying time meeting with suppliers, and roughly 75% say they prefer a rep-free experience. Buyers tell reps little, so the specific requirements they do voice on calls are valuable evidence.
4. Single-reason fields. Real losses have a primary reason plus contributing ones. The CRM forces you to pick one.
5. Counting deals instead of dollars. Ten small losses to price can outnumber three large losses to a missing feature, even when the three large losses cost far more revenue.

Underneath all five is a sixth problem: there is no feedback loop. When someone does spot a pattern, product hears an anecdote ('a few people asked for X') instead of evidence, and the anecdote loses the roadmap argument.

The Method: 6 Steps to Run Lost Deal Analysis From Sales Calls

You can run a complete lost deal analysis with a spreadsheet, your call transcripts or recordings, and a few focused hours. No tool is required for any of the steps below.

Prerequisite: this only works if your calls are recorded. If they aren't yet, start with our guide to recording Zoom, Meet and Teams calls locally. Always tell participants you are recording and get consent where the law requires it. Recording laws vary by jurisdiction.

Step 1: Pull the set

Export every closed-lost opportunity from the last 90 days with these fields:

• Deal ID
• Amount
• Stage reached
• Close date
• CRM loss reason
• Owner

Then gather every recorded call for each deal, including discovery calls, demos and security reviews, not only the last call. Mark any deal with no recordings as a coverage gap. Don't guess a reason for it.

Step 2: Extract the evidence per deal

For each deal, log four things. Each entry needs a verbatim quote and a call date or timestamp.

• Stated reasons: what the buyer said about why they hesitated or left.
• Objections: what was raised, and whether anyone answered it.
• Competitor mentions: who else was in the evaluation, and any existing contracts.
• Missing features or requirements: listen for phrases like 'we need,' 'does it support' and 'our security team requires.'

The rule is simple: no quote, no entry. Also note who raised each issue. Requirements voiced by IT, security or procurement reviewers can veto a deal even when the champion seemed happy.

Step 3: Separate stated from observed reasons

Keep the CRM reason in its own column. In a second column, record the observed primary reason, which is the blocker with the strongest call evidence, plus up to two contributing reasons. Use four tests:

1. Was it raised more than once?
2. Was it raised before pricing was discussed?
3. Did the deal stall or go dark soon after it came up?
4. Was it never resolved?

A blocker that passes three or four of these tests outranks the CRM label.

Step 4: Group and sum

Build a shared taxonomy of six to ten buckets. A good starting set:

• Missing capability
• Security/compliance
• True budget or price
• Lost to competitor
• No decision/status quo
• Timing or champion change

Tag each deal's observed primary reason, then sum pipeline dollars and deal count for each bucket. Sort by dollars first. Show deal count next to it so one large deal can't dominate the roadmap on its own.

Step 5: Send product gaps to product with evidence

For each capability gap, write one page that includes:

• The gap, stated in one sentence
• The deals affected and the total pipeline
• The stage each deal reached
• Three to five of the strongest buyer quotes, with dates
• What the buyer said would have changed the outcome

Then sort every reason by who can fix it:

• Product: missing capability.
• Sales: unanswered objections, slow follow-up, weak business case or poor qualification.
• Neither: budget freeze, acquisition or a departed champion.

'It's on the roadmap' does not answer an objection. If no dated commitment followed, log it as a Sales-fixable failure as well as a Product gap. If you find many of these broken promises, read our guide on tracking commitments made on sales calls.

Step 6: Re-check after the next quarter

Ninety days later, re-run the same steps with the same taxonomy. Then:

• Compare pipeline dollars per bucket against the baseline.
• Check whether shipped fixes reduced that reason's share of lost pipeline.
• Re-contact accounts that were lost to a gap you have since closed.

The lost deal analysis worksheet

Time estimate: plan on 20 to 40 minutes per deal if you skim transcripts instead of listening to full recordings. A 25-deal quarter works out to roughly 8 to 17 hours of review. That number matters for the threshold section below.

Worked Example: 38 Lost Deals, 22 'Pricing' Labels and a $1.1M SSO Gap

This is an illustrative scenario built to show how the method works. It is not a customer case study.

A B2B SaaS team closes the quarter with 38 lost deals worth $2.6M in pipeline, and the CRM lists 'pricing' on 22 of them. Leadership is debating a 15% discount to recover win rate.

What the calls showed

In Steps 2 and 3, the team reviewed transcripts for all 38 deals. They found that 14 deals, worth $1.1M, raised SSO/SAML as a requirement. The request usually came from the buyer's IT or security reviewer.

• Eleven of those 14 deals were coded 'pricing' in the CRM.
• In most of them, SSO came up before price was discussed.
• The rep's answer was usually 'it's on the roadmap,' with no date attached.

For background, single sign-on lets users log into many applications with one set of credentials. SAML 2.0, an OASIS standard since 2005, is the protocol behind most enterprise SSO requirements, alongside OIDC. Many security teams won't approve a new SaaS vendor without it. Gartner reports that 77% of B2B buyers describe their latest purchase as very complex or difficult. Buying committees with IT, security and finance reviewers are exactly where requirements like SSO surface mid-cycle and then get flattened into 'pricing' at close.

Stated vs observed loss reasons

The reframe

True price-driven losses were 8 deals and $420K, not 22 deals. A blanket discount would have cost the team in two ways. First, it would have given up margin on every future won deal. Second, it would still have left the $1.1M SSO problem untouched. Price is a large profit lever. A McKinsey analysis published in Harvard Business Review by Marn and Rosiello found that a 1% price improvement produced roughly an 11% gain in operating profit for the average company studied. Cutting price based on mislabeled data moves that lever in the wrong direction.

The decision

Product received a one-page evidence pack covering 14 deals and $1.1M, with five dated quotes such as 'our security team won't approve anything without SAML.' Product moved SSO into the next quarter's roadmap, ahead of a feature that had only anecdotal support.

Sales made three changes:

• Reps stopped answering 'it's on the roadmap.'
• Reps started qualifying for SSO in discovery.
• Buyers who needed SSO got a dated commitment instead of a vague promise.

The discount proposal was dropped.

The re-check plan

Next quarter, the team will track the share of lost pipeline tagged 'missing SSO.' Once SSO ships, they will re-open the 14 accounts with a direct note. The outcome isn't claimed yet. These are the measures that will show whether the fix worked.

The CRM told the team to cut price. The calls told them to build SSO.

Common Mistakes in Lost Deal Analysis

Most failed lost deal analyses fail for the same reason: the team treats opinion as evidence. Watch for these mistakes:

• Analyzing only the final call. The blocker usually appears mid-cycle, so review every call on the deal.
• Trusting the stated reason as data. Keep it as a column to compare against. It is not the answer.
• Coding without quotes. A tag without verbatim evidence is opinion again, and it will lose the argument with product.
• Counting deals instead of pipeline dollars. Deal counts hide large losses behind many small ones.
• Changing the taxonomy every quarter. If the buckets keep moving, you can't see trends.
• Ignoring 'no decision.' In The JOLT Effect, Dixon and McKenna analyzed more than 2.5 million recorded sales conversations. They found that 40 to 60% of deals where buyers expressed purchase intent ended in no decision, and about 56% of those losses came from buyer indecision rather than a preference for the status quo. These losses usually point to a weak business case, not a product gap.
• Skipping won deals entirely. If won deals mention the same gap, it may not be what actually caused the losses. Sample a few wins as a control.
• Overreacting to tiny samples. Three deals is a signal to watch. It is not enough to justify a roadmap change.

Some failure modes no tool can fix, BuildBetter included:

• Deals lost before any call was recorded
• Buyers who ghosted and never said why
• Reps who don't log calls to the right opportunity
• Decisions made in meetings you were never in, which only a post-loss buyer interview can uncover
• Recordings made without proper consent

When the Spreadsheet Stops Working (and What to Use Instead)

A spreadsheet works well for a handful of lost deals a quarter, but past roughly 20 it turns into days of work. At 20 to 40 minutes per deal, with several calls per deal, the review gets skipped. Nobody re-listens to the calls, quotes stop being pulled, and the analysis goes back to CRM dropdowns.

Other signs you've outgrown the manual version:

• Calls are scattered across Zoom, Meet and Teams.
• You need the original buyer quote months later, when the gap finally ships.
• Product asks 'how many deals?' and nobody can answer without a new review.
• The quarterly re-check never happens.

1. BuildBetter for Sales: lost deal intelligence

BuildBetter shows why deals stalled and which gaps product can address. It also provides deal intelligence across every call, including buying signals, objections and commitments. The sales-to-product handoff carries the full call context, so product sees the requests behind each deal. Signals structures every conversation by type, severity and business impact. Tracked Objects keeps commitments like 'we'll get back to you on SSO' linked to evidence.

• Integrations: Zoom, Google Meet, Teams, Gong (import), Slack, HubSpot, Salesforce, Gmail, Outlook, Zendesk and Intercom.
• Pricing: based on data ingestion, with unlimited seats, so sales, product and CS all see the same evidence.
• Compliance: SOC 2 Type II, HIPAA and GDPR.

Win the deal. Keep the promise.

'Looking back, we were lucky to go with BuildBetter. A lot of other tools were extremely rigid.' — James Wallis, Brex

2. BB Recorder: free call capture

BB Recorder is free, needs no account and works 100% offline. It records Zoom, Google Meet and Teams locally with no bot on the call. If you prefer a visible bot, you can switch to one mid-call. Transcription happens on your device.

• Platforms: Mac, Windows (including ARM), iPhone and Android (beta).
• Workspace upload: you can upload recordings to a BuildBetter workspace later.
• Download: buildbetter.ai/recorder.

BB Recorder does not collect consent for you. Tell participants you are recording and follow the laws that apply to you. A common setup is to capture calls for free with BB Recorder and use BuildBetter to turn them into lost deal intelligence, call prep and follow-ups. If you already pay for a call-recording platform and are weighing costs, see BB Recorder vs Gong.

No-budget option

If you have no budget, export your transcripts and paste them, one deal at a time, into a general-purpose LLM with the Step 2 extraction instructions. Keep the spreadsheet as the system of record, and check every quote against the original transcript before you use it.

A caveat on fit: if your main need is post-loss buyer interviews or large survey programs, a dedicated research process fits better. Call-based analysis only sees what was said on calls.

FAQ: Lost Deal Analysis From Sales Calls

How many lost deals do you need for a useful analysis?

About 15 to 20 lost deals in a quarter is enough to see patterns in pipeline dollars. With fewer, treat findings as signals to watch rather than roadmap decisions, or combine two quarters using the same taxonomy.

How far back should a lost deal analysis go?

Use the last 90 days of closed-lost deals. That window is recent enough that your product and market haven't changed much, and long enough to gather a meaningful sample. Re-run it every quarter with the same taxonomy so trends stay comparable.

Why does every closed-lost deal say 'pricing'?

'Pricing' is the easiest dropdown option for a rep to defend and the most polite exit for a buyer. It needs no evidence and ends the conversation. The underlying blocker, often a missing feature, a security requirement or a competitor, usually appears earlier and more specifically in the call record.

What is the difference between stated and observed loss reasons?

A stated reason is what the CRM field or the buyer says at the end. An observed reason is the blocker the call evidence supports: it was raised repeatedly, stayed unresolved and came before the stall. Keep both in separate columns so the gap between them is visible.

What is the difference between lost deal analysis and win-loss analysis?

Win-loss analysis compares won and lost deals, often through post-decision buyer interviews. Lost deal analysis from calls uses conversations you already recorded. It is faster and cheaper, but it can't see decisions made off-call. Many teams run call analysis continuously and use buyer interviews selectively for large losses.

How do you get product to act on lost deal findings?

Send each gap as a single page with the number of deals, the total pipeline lost and three to five dated buyer quotes. Evidence weighted in dollars wins roadmap arguments that anecdotes lose.

Is it legal to record sales calls for this analysis?

It depends on the jurisdiction. Some places require every party to consent. Always disclose that you are recording and get consent where required. Recording software alone does not make you compliant.

Sources and Further Reading

• Telephone call recording laws (Wikipedia): an overview of consent requirements by jurisdiction
• Single sign-on (Wikipedia): background for the worked example
• Security Assertion Markup Language (Wikipedia): the standard behind most enterprise SSO requirements
• Matthew Dixon and Ted McKenna, The JOLT Effect (Portfolio/Penguin, 2022)
• Michael Marn and Robert Rosiello, 'Managing Price, Gaining Profit,' Harvard Business Review (1992)
• Gartner B2B buying journey research
• Related BuildBetter guides: How to track commitments made on sales calls and How to record Zoom, Meet and Teams calls locally

Make Churn Optional

Your buyers already told you why they didn't sign. BuildBetter turns every recorded call into evidence-backed loss reasons, product gaps weighted by pipeline, and follow-ups to the accounts you lost once you ship the fix.

Make churn optional. Book a demo.