Win/loss analysis means interviewing buyers after deals are won, lost or end with no purchase, to learn what decided each. Run it by phone. Use an interviewer with no stake in the sale, call each buyer within four weeks of the decision, and record only with consent.
In This Article
Win/loss analysis is a set of interviews with buyers after a purchase decision, asking what decided it. The CRM records something else. A rep picks a loss reason from a dropdown, price or timing or no decision, and the record closes. Nobody asks the buyer.
The market research outsourcing service page describes full win/loss studies run by third-party interviewers, from scoping to the report. Each term below keeps one meaning.
- Win/loss analysis: interviews with buyers after a purchase decision, covering wins, losses and no-decision losses, to learn what decided each.
- Study: one research project with its own sample, guide and report; a win/loss study is one round of win/loss interviews.
- Program: a recurring series of win/loss studies, run each quarter on the same interview guide.
- Interview: one phone conversation with one buyer.
- Survey: a questionnaire sent to many respondents; here it names only the third-party research cited and the topics of sibling guides.
- Decision: the buyer’s final choice of the sponsor, a rival or no purchase at all.
- No-decision loss: a deal whose decision was no purchase at all, named for the sales term.
- Loss reason: the CRM code a rep selects.
- Sponsor: the company whose deals are studied.
- Sample frame: the list of closed deals a study draws from, such as a weekly CRM export of the past two weeks’ closed deals.
- Incidence: the share of screened contacts who qualify for a study, before any interview starts.
- Sample: the deals chosen from the frame.
- Interview guide: the written questions and probes every interviewer asks in the same order on every call.
- Probe: a follow-up question or prompt that tests or deepens the buyer’s answer before the next question.
- Complete (n): an interview that clears the complete threshold the study sets before fielding; only completes count toward n.
- Fielding: the weeks interviewers spend calling the sample.
- Disposition: the final AAPOR status of each sampled deal, such as complete, partial, refusal or non-contact.
- Verbatim: an answer in the buyer’s own words.
- Coding: sorting verbatims into countable categories.
- Decision driver: a category of reason, coded from verbatims, that counts as a driver once it recurs across interviews.
What is win/loss analysis, and how is a sales win/loss analysis different from CRM loss reasons?
Win/loss analysis asks the people who made a buying decision what decided it, deal by deal. A sales win/loss analysis records the buyer’s account; CRM loss reasons record the rep’s account of the same deal.
The two accounts disagree for ordinary reasons. A rep closing a lost deal answers to a quota and a manager, and price is a loss reason nobody argues with. The buyer had a committee, a shortlist, a reference call that went badly and a contract clause legal would not sign. None of that reaches the dropdown.
What is a win/loss analysis for each team? The win/loss analysis definition holds for every team, but each team reads the findings for a different decision. In win/loss analysis, product management reads the findings for feature gaps and roadmap timing, and sales reads them for evaluation mistakes. Marketing reads them for claims buyers rejected.
The phone surveys versus online surveys guide explains why win/loss interviews sit on the phone side of the mode line. A web form cannot probe the answer. A first study is read once, while a program repeats the guide each quarter so shifting drivers show as a trend.
Should a third party run your win/loss analysis interviews, or can sales do them?
A third party should run the interviews when the deals are recent and the losses are the point. Buyers soften criticism for the rep who sold to them, and a neutral interviewer takes that relationship off the call.
The market research outsourcing page states the claim plainly: third-party interviewers get more candid answers than internal teams, especially on lost deals. The mechanism is social. A buyer who chose a competitor is talking to someone who lost a commission on that choice, so politeness wins. Price becomes the polite answer.
Buyers still use reps for information. In Gartner’s May 2026 survey release, 69% of 645 B2B buyers said they prefer to validate AI-generated insights with a sales rep. Gartner’s survey ran August to September 2025, and the release publishes no sample source, incidence or questionnaire length. Candor about a loss is a different matter.
Each account still earns a place:
- A rep debrief is fast, costs nothing extra and records what the rep saw in the room.
- A buyer interview by someone with no stake records what the buyer weighed after the rep left.
- Set side by side, deal by deal, the gap between the two accounts is a finding of its own.
An internal interviewer works when that person had no part in the deal, such as a researcher who never met the buyer. The test is whether the buyer owes politeness.
Who should you call for a B2B win/loss analysis, and how many interviews are enough?
Call the person who made the decision, within four weeks of it, across recent wins, losses and no-decision losses. A study of 20 to 30 completes, split across those outcomes, gives each side enough completes to show a pattern.
Start from the sample frame. A weekly CRM export of deals closed in the past two weeks, won and lost, above a minimum deal size, does the job. Then pick each buyer by role. Call the decision-maker or the evaluation lead, not the procurement contact who sent the regret email.
Timing protects the detail. The market research outsourcing page states that most phone studies are fully in-field within 5 to 7 business days of kickoff. That keeps week-one deals inside the window. The page publishes the figure with no split by audience, interview length, sample source, incidence or B2B and B2C.
Thirty completes is a working size. With 12 wins, 12 losses and 6 no-decision losses, a driver named in 5 of 12 losses is a pattern worth testing. A reason named once is an anecdote.
The counting standard is AAPOR’s Standard Definitions, the 2023 rulebook for classifying every case a study calls. Its example rule counts a complete at more than 80% of questions answered and a partial at 50% to 80%. The researcher fixes the rule before fielding. Put plainly: only completes count toward n.
For your study, this sets two operating facts. Report completes by outcome, never dials or partials, so 30 means 30 buyers who reached the complete threshold. Keep the final disposition of every sampled deal, because the buyers who never answered are a finding too.
The illustrative failure is ordinary. A study reports 30 interviews, but 9 of them ended at question 5, so the price and decision questions rest on 21 buyers. The report never says so.
Two safeguards follow from the standard. Ask any provider for the disposition table, completes and partials counted apart, before reading a single driver. Book the full appointment length, and treat a buyer who must leave before the complete threshold as a callback.
No-decision losses need their own row. Some deals that end with no purchase were never qualified, so check how each entered the pipeline. The B2B lead qualification service page describes live calls that verify budget, authority, need and timeline before a lead reaches a rep.
Table 1 lays out a 30-complete plan for a company losing ground to one competitor.
Segment | Completes | Who to call | What the segment answers |
|---|---|---|---|
Total | 30 | ||
Losses to the main competitor | 9 | Decision-maker or evaluation lead | lead What the competitor did that the sponsor did not |
Losses to other vendors | 3 | Decision-maker | Whether the loss pattern is one rival or the whole category |
No-decision losses | 6 | Evaluation lead | Whether the deal was real, and what stopped it |
Wins against the main competitor | 8
| Decision-maker | What held when the competitor was in the room |
Other wins | 4 | Evaluation lead | What buyers value without being prompted |
The split follows the question. A company losing to one rival weights that rival’s deals on both sides, because a win against it explains as much as a loss.
Tell us your research scope
Scope a 30-complete win/loss study in staff hours. Bring the decision it informs and the deals it covers, and the scope follows from them.
What are the win/loss analysis best practices for running the call itself?
A win/loss call runs by appointment, opens by naming who is calling and why, and lets the buyer tell the evaluation in order. The interviewer probes each answer once.
Five steps carry the call:
- The sponsor emails the buyer first, naming the interviewing firm and the purpose, so the call is expected.
- The interviewer books a time the buyer chose, then confirms it by email the day before.
- The call opens with the interviewer’s name and firm, the sponsor, the purpose, the length and the recording question. The call recording laws by state guide sets out which states require every party’s consent.
- The buyer tells the evaluation from first conversation to decision, and any rating the sponsor adds comes after that story.
- The interviewer closes on what would have changed the outcome, then thanks the buyer by name.
The order protects the data. A rating asked first anchors the story to the scale, so a sponsor that adds one puts it after the story. The B2B survey design guide for phone administration covers how rating items are worded and read aloud.
An incentive is a courtesy, not payment. A donation to a charity the buyer names sidesteps the gift limits companies and public agencies set for their staff.
Three more rule sets touch the call. The FCC’s consent revocation order, FCC 26-67, takes effect on 25 November 2026 and governs autodialed and prerecorded calls. A hand-dialed interview booked by email falls outside it, though a program that autodials wireless numbers follows it. State mini-TCPA laws regulate sales calls. A hand-dialed win/loss interview that never pitches stays outside them, and the state mini-TCPA reference guide tables each state’s rules. Research must also stay research. The Insights Association Code of Standards bars direct action toward a participant based on their answers without consent. No lost buyer goes to a rep.
Which win/loss analysis questions should an interview guide ask?
An interview guide asks about the trigger, the criteria, the shortlist, the evaluation and the decision, then ends on what would have changed it. Ten questions fit one appointment.
Wording beyond these ten belongs to the guide to survey scripts that get honest answers, which covers probes and rapport for every research call. Table 2 is the win/loss analysis template. Each question appears in the order it is asked, with its probe and the coding category it feeds.
Table 2 holds the ten questions, in order.
# | Question as read | Probe | Coding category |
|---|---|---|---|
1 | What made your team start looking at this purchase when it did? | it did? What was happening that month? | Trigger and timing |
2 | Before you saw any vendor, what did the purchase have to do? | Which of those was non-negotiable? | Must-have criteria |
3 | Which vendors did you consider, and how did each make the list? | Who was cut first, and why? | Competitive set |
4 | Walk me through the evaluation from the first call to the decision. | Where did it slow down? | Process and stalls |
5 | Who had a say, and whose view carried the decision? | What did that person care about? | Buying committee |
6 | What decided it in the end? | If you had to name one reason, which one? | Primary reason |
7 | How did each vendor’s sales team handle your questions? | Give me one moment that stood out. | Sales execution |
8 | How did the price compare with what you expected to pay? | Was price the reason, or the reason you gave? | gave? Price and value |
9 | What did the product you chose do that the others did not? | Did you see that in a demo or hear it? | Product gaps |
10 | What would the losing vendor have had to do to win? | Would that have been enough? | Actions |
Ask question 6 only after question 4. A buyer who has told the evaluation in order names the deciding reason from inside the story. Asked cold, the same buyer names the reason that is easiest to say out loud.
Question 8 earns its probe on losses. Price is the answer nobody questions, and the probe separates a real price gap from a value the buyer never saw.
Can you record a win/loss analysis call, and what consent does the law require?
Federal law allows recording a win/loss call when one party consents, and some states require consent from everyone on the call. Ask before the recording starts.
The rule is 18 U.S.C. § 2511(2)(d), the consent exception in the federal Wiretap Act. A party to the call, or a person with one party’s prior consent, can record it unless the purpose is criminal or tortious. Put plainly: one consent sets the federal floor. The stricter state sets the real line.
For your study, this sets three operating facts. The notice depends on where the buyer sits, so one scripted notice read to every buyer meets the strictest state. A transcription service or AI notetaker that keeps interview audio for its own use is a separate actor, outside the interviewer’s consent. Recording also makes verbatims checkable.
The pending case on that second fact is In re Otter.AI Privacy Litigation in the Northern District of California. In the court’s order of 13 August 2026, the judge let the federal Wiretap Act claim proceed past a motion to dismiss. That is a pleading-stage ruling.
The court accepted as true the allegation that Otter kept meeting audio to train its own systems. It held that committing a tort and seeking a profit are not mutually exclusive, so the one-party exception did not end the claim. Liability is not decided. Otter denies the allegations, and the case continues, so a win/loss team reads it as a warning about vendor terms.
Three safeguards follow from the rule. Read a recording notice at the start of every call, and record only after a spoken yes, wherever the buyer sits. Put the transcription vendor’s terms in writing: no training on interview audio, a retention limit and deletion on request. Strip names from verbatims before sharing. A buyer quoted to the rep who lost the deal will not take the next call from anyone.
This section describes federal law in general terms and is not legal advice for any particular study. Counsel should review notices and vendor terms.
What goes in a win/loss analysis report, and what does a working example look like?
A win/loss analysis report ranks the decision drivers by their count in wins and in losses, and quotes the buyers who named them. Each driver gets an owner.
Five parts make a report someone acts on:
- The decision the study informs, plus the completes it covered, counted by win, loss and no-decision loss.
- The drivers, ranked, with the count in wins and the count in losses side by side.
- Competitor by competitor, what each rival did in the deals it won from the sponsor.
- Verbatims under each driver, stripped of names.
- Actions, each with an owner in sales, product management or marketing, and a date to check again.
Coding is where the study earns its cost. Two people code the first ten verbatims separately, compare their categories and fix the list before coding the rest. Price then means one thing.
A win/loss analysis example makes this concrete. Take an illustrative loss a rep logged as price in the CRM, where the buyer’s verbatim says the vendor would not move on terms. The verbatim codes to negotiating flexibility, which sales leadership owns, rather than list price, which product and finance own.
Judge any provider’s report the same way. It should show completes by outcome, the fielding dates, who interviewed, and the verbatims under each driver it ranks. Without those, the report is opinion.
Findings about one rival feed a wider picture, and the guide to competitive intelligence by phone covers the other sources worth calling.
Should you hire win/loss analysis companies or run the interviews in-house?
Run win/loss in-house when a neutral person has the hours to schedule, interview and code; hire outside interviewers when nobody does. Price both options in hours per complete.
The hours are the hidden cost. Each complete needs scheduling attempts, the call itself, a notes or transcript check and coding, and a study multiplies all four. At an illustrative three hours per complete, a 30-complete study needs 90 staff hours before anyone writes the report.
When comparing win/loss analysis companies, compare quotes on hours or price per complete, then put four questions to each provider. Who conducts the interviews, and with what briefing? Are calls recorded with consent, and who keeps the audio once fielding ends and the study closes? Does the sponsor receive every verbatim? And who codes the verbatims into drivers, against which category list, checked by a second coder before the counts are final?
Win/loss works when findings reach an owner. A report nobody acts on costs the same 30 completes as one that changes the next quarter’s deals.
Frequently asked questions
What is win/loss analysis, and how is it different from the loss reasons my sales team logs in the CRM?
Win/loss analysis is a set of interviews with buyers after a purchase decision, covering wins, losses and no-decision losses. A CRM loss reason is the rep’s code. The interview records the buyer’s account in the buyer’s words, so it catches reasons a rep never heard, such as a failed reference call.
Should a third party conduct win/loss interviews, or can my own sales reps do them?
A third party, or an internal researcher with no part in the deal, should conduct win/loss interviews. Buyers soften criticism for their rep. Rep debriefs still help as a second account, and comparing the two versions deal by deal shows where the sales team’s picture is wrong.
How many win/loss interviews do I need, and how soon after the decision should I call?
A first win/loss study works at 20 to 30 completes, meaning interviews that clear the study’s threshold, split across wins, losses and no-decision losses. Call within four weeks of each decision. Choose the decision-maker or the evaluation lead for each deal, not whoever is easiest to reach.
What questions should I ask in a win/loss analysis interview with a lost buyer?
Win/loss analysis questions follow the purchase in order: the trigger, the criteria, the shortlist, the evaluation and the decision. Ask about price after the story. Then probe whether price was the reason or the reason given, and close on what the losing vendor needed to win.
Is it legal to record a win/loss interview with a buyer?
Recording a win/loss interview is lawful under the federal Wiretap Act when one party consents, though some states require everyone’s consent. A notice read to everyone on the call, followed by each person’s spoken yes, meets both. Check any transcription vendor’s terms too. A vendor that keeps the audio for its own use raises its own consent question.
Discuss your research project
