TL;DR
- The problem: The closed-lost reason field is filled in by the person who lost the deal, after the fact, from memory. That is a guess wearing a dropdown.
- The evidence: When Clozd compared 1,000 closed-lost CRM records against what buyers said in interviews, buyer and seller reasons agreed about 15 percent of the time.
- The method: Win-loss analysis is a research program, not a deal review. Five questions, asked by someone who was not on the deal, six to eight weeks after the decision.
- The output: Reason codes you can count, quotes you can hand to product, and a positioning correction you can ship next quarter.
- The cost: One owner, six interviews a quarter, and the willingness to hear that your message never landed.
Every CRM I have ever audited has the same field. It is called Loss Reason, or Closed Lost Reason, or Disposition. It has a dropdown with eight options and a free text box that almost nobody fills in. It is the single most quoted field in the revenue meeting and the least verified field in the company.
Ask a CMO why the quarter missed and you will hear the dropdown read back. Price. No budget. Lost to competitor. No decision. Four categories that sound like findings and are the absence of findings. None of them tells you what the buyer believed, what they compared you against, or which sentence in your pitch made them hesitate.
Teams build forecast models on that field and prioritize roadmaps with it. The only person who could tell you what actually happened was sitting in an inbox, willing to answer an email if someone asked.
The loss reason is written by the person with the strongest incentive to describe the loss as unavoidable. Price is not a reason. Price is what a buyer says when they do not want to explain the real reason to the person who lost.
Why the Loss Reason Field Is Fiction
Three mechanics make it unreliable, and none are the rep’s fault.
It is written late. The reason gets logged when the opportunity closes, often two to six weeks after the decision was made in a room the rep was not in. Memory reconstructs a story that makes the outcome sensible. That is why interview research exists as a discipline.
It is written by an interested party. A rep who writes lost on price is describing a market condition. A rep who writes we failed to prove ROI is describing their own execution. Both may be true. Only one requires a change in behavior, and the dropdown does not force the distinction.
It collapses a committee into one opinion. B2B deals are decided by a group, and the group rarely agrees on why. The economic buyer cared about risk. The practitioner cared about migration pain. The security reviewer killed it in week five and never spoke to your rep. One dropdown field cannot hold that, so it holds the simplest version instead.
Clozd put a number on the gap. The firm compared CRM loss reasons from 1,000 closed-lost deals against what buyers said when someone interviewed them, and found buyer and seller explanations aligned roughly 15 percent of the time. Clozd sells win-loss programs, so treat the number as directional. Directional is enough. If even half that gap is real, you run go-to-market on a field that is wrong more often than right.
This is the same failure I wrote about in the customer intelligence loop: the CRM records what your team did, not what your buyer thought. The activity is logged. The reasoning is not. And the reasoning is the part that would change your next quarter.
Win-Loss Analysis Is Research, Not a Deal Review
Most teams think they already do this. What they do is a deal review: the rep presents the timeline, the room agrees the buyer was difficult, and everyone leaves with the story they arrived with. That is a post-mortem of your process, not research.
Consider the timing. 6sense’s 2025 buyer research, a survey of roughly 4,000 buyers, puts 61 percent of the journey ahead of first seller contact. By the time a rep is in the room, most of the decision is already shaped. The interview is the only route back into the part you missed.
Four properties separate it from a post-mortem.
The buyer is the source. Not the rep, not the call recording, not the champion who stopped replying. The person who signed or declined, on a call, in their own words.
Someone neutral runs it. Buyers will not tell a rep that the demo felt generic, because it is rude and there is no upside. They will tell a researcher who is not attached to the outcome, especially when told the rep will never hear their name on the quote.
It samples, it does not summarize. Ten interviews across two quarters is a sample. One angry loss is an anecdote. The discipline is in the count.
It codes before it concludes. Every interview gets tagged against a fixed set of reason codes, blind to the CRM field, and then you compare. This is the step that turns interviews into a dataset instead of a folder of quotes.
The Five Questions That Crack a Deal Open
You do not need a 40-question instrument. You need five questions and the discipline to shut up after asking them. These are the ones I use.
| Question | What it surfaces |
|---|---|
| Walk me through the moment you decided to look at this at all. | The trigger event. Most teams never learn it, and it is usually a change in headcount, a failed audit, or a new boss. |
| Who else was involved, and what did each of them need to be true? | The buying committee and the real veto holder. Frequently not the person your rep was talking to. |
| What did you compare us against, including doing nothing? | The true competitive set. Doing nothing is the most common alternative, and it never appears in competitive analysis. |
| What was the one thing that made you hesitate? | The objection that never reached your rep. This is where messaging failures show up. |
| If you had to explain this decision to your board in one sentence, what would you say? | The buyer’s own positioning language. Use it verbatim in your next deck. |
The fifth question is the one I would keep. Buyers compress decisions into sentences, and those sentences are raw material for positioning. When a buyer says they chose the other vendor because it could get them live before the fiscal year closed, you did not lose on features. You lost on time to value, and your website probably does not mention it once.
How to Run It In-House in Four Steps
You do not need a vendor or a research budget. You need six weeks and one owner.
Take every closed-lost and closed-won opportunity from the last two quarters. Sample 20. Write down the CRM reason for each, then set that column aside and do not look at it again until coding. If you read the field first, you will ask leading questions and confirm the story you already have.
Twenty-five minutes, recorded with permission, run by someone who was not on the deal. Marketing, product marketing, or an ops analyst works fine. The script is five questions. The skill is silence. Every time you explain yourself, you lose a data point.
Six to eight codes, agreed before the first call. Fit, timing, price and budget, missing capability, risk and trust, internal politics, no decision. Two people code the same three interviews independently to check agreement. Then compare your codes to the CRM field and count the mismatches.
A finding with no owner is a slide. Messaging findings go to positioning. Objection findings go to sales enablement. Capability findings go to product. Committee findings go to the account plan. Put a name and a date against each one, or the program dies at the readout.
Six to eight weeks after the decision is the sweet spot. Sooner and the buyer is still inside the emotion. Later and they have rationalized it, which is the state your rep was in when they filled in the dropdown.
What I Actually Think
Here is the part that took me years to accept. The loss reason is not a data quality problem. It is a status problem. Nobody wants to be the person who lost because the message was wrong, so the system rewards the reason requiring no change. Price is the most popular entry because it is the only one where nothing could have been done.
I stopped asking reps why we lost. I ask them what they think the buyer would say, write it down, then go ask the buyer. The gap between those two answers is the most valuable document in the company. It tells you exactly where your story stops working.
Most teams also run win-loss too late in their maturity. The instinct is to wait for a research budget, a product marketing hire, and a clean CRM. That order is backwards. The interviews teach you what the CRM should have captured. Run six now with a spreadsheet and a shared calendar, then let the taxonomy emerge from what buyers say.
And I will say the unpopular part. The champion is not the buyer. The person who loved your demo, pushed internally, then went quiet is the least reliable subject you have, because they are managing their own embarrassment. The useful conversations are with the person who said no and the buyer who chose the status quo. Both will tell you the truth in fewer sentences than any analyst report.
Where the Findings Actually Go
A win-loss program fails in one of two ways. It produces no findings, because the questions were soft. Or it produces findings that change nothing, because nobody owned the routing. The second failure is more common and more expensive.
I route findings into four buckets, and each bucket has a different owner and a different artifact.
Positioning. If buyers describe the value differently than you do, your homepage is wrong, not the buyer. The artifact is a revised H1 and a revised proof section, shipped in the same quarter. Buyer language goes in verbatim. This is the principle behind treating customer intelligence as a system, not a dashboard: the point is not to know more, it is to change what you say.
Enablement. If the hesitation happens in the middle of the deal, the rep needs a specific asset, not a training session. The artifact is one objection page per recurring hesitation, with the exact language the winning buyers used.
Product and roadmap. Missing capability findings are the only ones product teams will act on, and they will only act if the count is visible. Ten interviews, four mentions of the same integration gap, is a signal. One mention is a request.
Qualification. If lost deals share a firmographic pattern, the problem is upstream in targeting. Most pipelines are not short on leads, they are short on buyers who were ever going to say yes. The dark funnel work makes the same point from the other side. The decision formed before you were in the room, with people who were never in your CRM.
One governance note. The person who owns the program should not own the number. If the VP of Sales runs win-loss, findings bend toward execution stories. If product marketing runs it, findings bend toward messaging. Either is defensible as long as you name the bias in the readout.
Start With Six
Six interviews is a quarter of work at two hours each, writeup included. Small enough to start this month, large enough to see a pattern. Pick six closed-lost deals from the last 90 days. Have someone neutral send the request. Ask five questions. Code the answers before you look at the dropdown.
Then put the CRM field and the interview answers side by side on one page. That page does more for your go-to-market than another dashboard, because it replaces a guess with a sentence a buyer actually said. The first time you show it, someone will get defensive. That reaction is the finding.
Your CRM loss reason is an educated guess, written late, by the person who lost. Win-loss analysis replaces it with the buyer’s own sentence. Five questions, a neutral interviewer, six to eight weeks after the decision, coded against a fixed list, routed to a named owner. It costs a quarter of one person’s attention and corrects the story your revenue motion is built on.
