How do you run a win-loss review? A five-question method for small teams

Why not just use the loss reason in your CRM?

Because the CRM field usually records what the rep believes. A rep who lost on a call that went quiet picks "price" or "timing" because it's the easiest option in the list, and nobody asks the buyer. That's our view, and the research-firm guidance we read points the same way: Drive Research notes that buyers tend to hold back the full story when the person asking is the rep who lost, and it recommends a neutral interviewer.

So a CRM field is a useful index, and a buyer interview is the evidence. Before you start, it helps to know how you count wins and losses; see what a good B2B win rate is.

Who should run the interview, and when?

• Who: anyone except the rep who ran the deal. A founder, a teammate on another account, or a paid outside interviewer. The point is that the buyer has no reason to be polite.
• When: soon after the decision. Chief's guide suggests 2 to 4 weeks, and Unkover's says buyers rationalize their choice after about 30 days; both are vendor guides, not studies. Drive Research says the same in general terms: soon, while the experience is fresh.
• Which deals: include wins and losses. Chief suggests a mix of wins, losses and no-decisions, spread across deal sizes, industries and reps. Deals that ended with nobody buying are worth a call too; the five types and how to tag them are in why B2B deals end in no decision.

What are the five questions to ask?

These are our questions, written to be open-ended and to avoid fishing for compliments. Keep each interview to 15 to 20 minutes.

Swipe to see more →

Don't ask "would you recommend us?" or "was our price fair?". They invite a yes or a polite no. Ask follow-ups such as "what do you mean by that?" instead of defending a point.

How many interviews do you need?

There's no agreed number. Chief suggests starting with 6 to 10 a quarter and says 20 or more give a statistically meaningful picture, while Unkover recommends 15 to 20 a quarter. Both are vendor claims. For a team closing a few deals a month, our suggestion:

1. Five interviews in the first month, covering at least two lost deals and two wins.
2. Ten by the end of the quarter.
3. Look for patterns only when the same reason shows up in at least three interviews. That's our rule of thumb, not a statistical test. Five interviews can't give you a percentage.

An invitation to send, invented for illustration:

Expect many buyers not to reply. Unkover cites a 10% to 30% response rate for recruiting, which is a vendor figure, so invite more deals than you need.

How do you record and tag what you hear?

One row per interview, in a spreadsheet:

Swipe to see more →

Keep the buyer's own words in the cells and tag afterwards, so that your tags don't shape what you hear. Chief's guide uses categories like sales execution, product fit and pricing, so you can adapt them.

How do you turn the answers into changes?

Hold a 30-minute review once a month. For each tag that appears in three or more interviews, pick one change and one owner:

• ICP fit: tighten who you target. See how to qualify with BANT or MEDDIC.
• Missing stakeholder: add a "who else has to agree?" question to your first call. See how many people are in a B2B buying committee.
• Proof or payback: put a customer result or the payback math earlier in the deck. See the pitch deck guide.
• Outreach relevance: change who you contact and why. See cold emails that work.

Change one thing at a time, so that you can tell what moved your win rate.

When should you skip each step?

• An outside interviewer: skip it if you can't afford one. A teammate who wasn't on the deal is the next best option, but don't use the rep or their manager.
• Win interviews: skip them for the first batch if you have very few wins. In our view losses teach more per interview, and you can add wins later to learn what to repeat.
• A monthly review meeting: skip it if you have fewer than about five interviews. Read the notes alone and wait.
• Scoring the buyer's answers: skip any numeric score on five interviews. The sample is too small, and the quote is more useful than a percentage.
• Recording calls: skip it unless the buyer agrees at the start of the call. Consent rules for recording differ by country and US state, so check the ones that apply to you, or take notes instead.