How to Run a Win/Loss Analysis: A Repeatable Process

To run a win/loss analysis, interview buyers 2 to 6 weeks after a deal closes, ask them to walk through how they made the decision, code their answers into a small set of themes, and review the results every quarter with a named owner for each action. Interview both wins and losses, use someone who did not work the deal, and weigh themes by revenue, not by how often a rep mentions them.

The rest of this article gives you the full process: who to talk to, the questions that work, how to synthesize the answers, and how to make sure something changes afterward.

Why your CRM loss reasons are not enough

Most teams already have a "closed-lost reason" field. It is usually wrong. Reps pick the option that is fastest to select and least personally uncomfortable. "Price" and "went with competitor" dominate because they point away from anything the rep or the company could have done differently.

Buyers, when asked directly, often tell a different story. Price is frequently a stand-in for "I could not see enough value to justify this," which is a positioning or sales-process problem, not a pricing problem.

Here is how the main sources of deal feedback compare:

Use all four if you can, but treat the buyer interview as the anchor. The others help you check it.

What you need before you start

You do not need software or a budget. You need four things.

• A defined scope. Pick one segment, product line, or deal size band for your first round. Mixing enterprise and self-serve deals muddies every pattern.
• A deal list. Pull closed deals from the last 60 to 90 days with contact names, deal size, competitor (if known), and the rep.
• An interviewer. Someone neutral. In a team of 10 to 20 people, this is often a product marketer, an ops lead, or a founder who was not in the deal.
• A place to store notes. A spreadsheet is fine. One row per interview, one column per theme.

If you have not yet defined who you are trying to win, do that first. A good ideal customer profile lets you tag each deal as in-profile or out-of-profile, which changes how you read the results.

Step-by-step: how to run a win/loss analysis

Step 1: Choose the deals to sample

Select 12 to 16 deals per quarter. Include both outcomes, and include a deliberate mix:

• 6 to 8 losses, including at least 2 to a competitor and at least 2 to "no decision"
• 4 to 6 wins, including at least 1 you almost lost
• Deals above your median size, since they carry more revenue weight

Do not hand-pick only the deals you are curious about. Pull them by rule (for example, every closed deal over $10,000, then a random sample of the rest) so you do not confirm what you already believe.

Step 2: Request the interview the right way

Send a short, honest note within a week of the close. Say you are not trying to resell, you want to learn, and the call takes 20 minutes. Offer a $25 gift card if your buyers are small businesses and it fits your culture. Expect roughly 25 to 40 percent of lost buyers and 50 percent or more of won buyers to agree.

Keep the request from the interviewer, not the rep. A message from a neutral person gets more candid answers.

Step 3: Run the interview

Aim for 20 to 30 minutes. Record it with permission. Your job is to get the buyer to tell the story of the decision in order, not to defend your product.

Use this question outline:

Context

• What was happening in your business that made you start looking?
• What was the cost of doing nothing?

Process

• Who was involved in the decision, and what did each person care about?
• Which options did you consider, and how did you find them?

Decision

• What were the top two or three things that decided it?
• Was there a moment when you became more or less confident in us?
• What nearly changed your mind?

Experience

• How did our sales process compare to the others?
• Was anything unclear, missing, or surprising?

Close

• If you could change one thing about our offer, what would it be?
• What advice would you give us?

Two rules matter more than the questions. First, ask "what else?" after every answer at least once, because the second reason is often the real one. Second, stay quiet after the buyer finishes speaking. Silence gets you the unrehearsed part.

Step 4: Code the interviews within 48 hours

Do not wait until you have all the interviews. After each call, fill in one spreadsheet row:

• Deal size, segment, competitor, outcome
• The buyer's top three decision factors, in their words
• One verbatim quote that captures the decision
• Tags from a short list: price, product gap, integration, implementation risk, trust or brand, sales experience, timing, internal politics, status quo

Keep your tag list under 12 items. If you invent a new tag for every interview, you will have no patterns.

Step 5: Synthesize by weight, not by count

At the end of the quarter, sort tags by two measures: how many buyers mentioned the theme as a top-three factor, and how much revenue sat behind those deals. A theme cited in 3 deals worth $40,000 each beats a theme cited in 5 deals worth $5,000 each.

Then do the comparison that makes win/loss work: line up the top decision factors for wins against the top factors for losses. Themes that appear only in losses are your leaks. Themes that appear only in wins are your real strengths, which may differ from what your marketing says. If you want help sharpening those strengths into messaging, the positioning statement guide is the natural next step.

Finally, compare the buyer's story with your CRM story. Where they disagree, you have found a blind spot.

Step 6: Write a one-page readout

The readout has four parts:

1. Sample and outcome. How many interviews, win rate, and revenue covered.
2. Top three themes. Each with a number, a quote, and revenue impact.
3. Gap between belief and evidence. What the team assumed versus what buyers said.
4. Actions. Each with an owner, a due date, and a measure.

One page. If it needs more, you are reporting instead of deciding.

Step 7: Assign actions and follow up

Present the readout in a 45-minute meeting with sales, product, and marketing in the room. End the meeting only when each action has one named owner. Record the decisions in a decision log so that in two quarters you can check whether the change moved the numbers.

A worked example

A 14-person B2B software company sold a scheduling tool for field service teams, with an average contract value of about $18,000 a year. In one quarter it closed 96 qualified opportunities: 22 wins (a 23 percent win rate) and 74 losses.

The CRM told a simple story. Reps tagged 58 percent of losses as "price." The leadership team was ready to discount and had a pricing change drafted.

Before acting, the head of marketing ran 16 loss interviews and 5 win interviews over five weeks, about 10 hours of calls in total. The coded results:

• Price was a top-three factor in 7 of the 16 loss interviews, but the buyer's primary reason in only 3.
• Implementation risk ("we could not tell how long setup would take or who would do it") was the primary reason in 9 of the 16, and these were the larger deals, averaging $26,000.
• A missing integration with a popular accounting tool was the primary reason in 4.

Wins told the matching story. Four of the five winners said a clear onboarding plan shown in the second call was the deciding moment.

The team dropped the discount plan. Instead it built a one-page implementation plan with a 30-day go-live commitment, trained reps to present it in the second call, and moved the accounting integration up the roadmap. The readout named an owner for each: the sales lead for the plan, the product lead for the integration.

Two quarters later, win rate on qualified mid-market deals rose from 23 percent to 29 percent. On roughly 96 deals a quarter, that is about 6 extra wins, or around $108,000 in additional annual contract value per quarter. The discount they nearly shipped would have cut revenue on every deal, including ones they were already winning.

Cadence: how often to run it

Win/loss is a habit, not a project. The cadence that works for most small teams:

• Weekly: The interviewer books and runs 1 to 3 interviews, and codes each within 48 hours. About 1 hour a week.
• Monthly: A 15-minute check-in with sales leadership to flag urgent patterns, such as a new competitor appearing in 4 of 10 interviews.
• Quarterly: The one-page readout and action review. Retire actions that are done and re-test the ones that are not moving the numbers.
• Yearly: Revisit the sampling rules, tag list, and question set. Check whether the segments you sample still match the market you target. This pairs well with your competitive analysis refresh.

If you close fewer than 10 deals a month, drop to 3 or 4 interviews monthly and a lighter quarterly summary. A small, steady sample beats a big burst once a year.

Common mistakes and how to avoid them

Letting the salesperson run the interview

This is the most common failure. The buyer does not want to insult someone they just spoke to for weeks, so they say "it came down to price" and move on. The rep, meanwhile, hears what they expect to hear.

Fix: Use someone outside the deal. If you are a solo founder, hire a freelancer for a few hundred dollars per batch of interviews, or ask a trusted peer to run them while you read the transcripts.

Interviewing only losses

Losses feel urgent, so teams skip wins. You then learn what to fix but not what to protect. You may also "fix" something that your winning customers actually love.

Fix: Keep wins at 30 to 40 percent of your sample, and always ask winners what almost stopped them.

Taking the first answer at face value

"Price" is the most common first answer and the least reliable one. It is polite, fast, and hard to argue with.

Fix: Follow up with "Compared to what?" and "What would have made the price feel worth it?" Those two questions often expose the real issue in under a minute. Also be aware of your own bias when you read the notes: confirmation bias and similar traps make it easy to highlight only the quotes that support your existing plan.

Producing a report nobody owns

A readout without owners becomes a slide deck that people nod at. Nothing changes, the next round surfaces the same themes, and the team concludes win/loss "doesn't work."

Fix: No action leaves the meeting without one owner, a date, and a measure. Review last quarter's actions first at every readout.

Drawing conclusions from too few interviews

Three interviews feel like a pattern. They are an anecdote.

Fix: Do not change strategy on fewer than 8 interviews unless the same reason appears in nearly every one. Below that, treat findings as hypotheses to test in the next round.

Key takeaways

• Buyer interviews, not CRM loss reasons, are the anchor of a good win/loss analysis. Use the CRM for volume and the interviews for truth.
• Have someone who did not work the deal run the call, 2 to 6 weeks after close, and ask the buyer to walk through the decision in order.
• Sample both wins and losses, aim for 10 to 15 interviews per quarter, and code each within 48 hours using a tag list of 12 items or fewer.
• Weigh themes by revenue behind them, and compare win drivers against loss drivers to find real leaks and real strengths.
• Finish every quarter with a one-page readout where each action has a named owner, a date, and a measure.
• Run the process on a steady cadence. A small weekly habit produces better decisions than a large annual project.