Filter your CRM to closed-lost and read the reason column. For many early-stage teams it holds a short list of phrases: price, went with a competitor, no decision, went silent. Those phrases feel like answers. They are what a founder typed in a few minutes after a deal died, from one side of the table, and they quietly decide what gets fixed next: the pricing page, the demo, the product roadmap.
If you run your own deals, this is worth an hour of your attention. The reason you record is a hypothesis, and acting on an unconfirmed hypothesis is expensive. This article argues for a different practice: treat every loss reason as unverified until the buyer has confirmed it, and have the founder collect that confirmation personally, within two weeks of every decision.
The reasons you record most are not reasons
Lightspeed Venture Partners surveyed 154 venture-backed sales teams on why deals fail to close. As SaaStr's Jason Lemkin summarised the results, the number one recorded reason was "no action" and the third was "went silent". His verdict: "These aren't reasons. These are non-reasons." A buyer who took no action had a reason for it. Nobody wrote it down.
The gap between what sellers record and what buyers say is large. Corporate Visions, a firm that sells win-loss services, analysed more than 100,000 B2B purchase decisions across 500 companies and reports that sellers and buyers give different reasons for a lost deal 50 to 70 percent of the time. The same analysis says 53 percent of buyers believe the losing vendor could have won with a fixable change, and that in 10 percent of deals marked lost, the buyer was still weighing options. Treat those figures as that vendor's findings, not settled benchmarks. The direction is still hard to ignore.
One reasonable reading is that the loss-reason dropdown measures how the seller felt about the deal, not why the buyer decided. Buyers who pass often volunteer only the polite version, and a seller who just lost a deal is not looking for the harder one. The field gets filled with whichever answer closes the loop fastest.
Fixing the wrong thing costs more than the lost deal
The damage is not the one bad entry. It is the decision built on top of it. "Lost on price" sends a founder to the pricing page. "Went silent" sends her to the follow-up sequence. Neither fix helps if the buyer never believed the project was urgent, or if the deal died in a security review nobody prepared for.
Reasons recorded in a CRM tend to point at things the seller can change this week. Reasons buyers give often point at things earlier in the system: whether the problem was urgent enough, whether the proof was believable, whether the deal could survive the buyer's own process. A founder who only reads the first set will keep repairing the last step of a motion that broke at the first.
Early wins make this harder to see
There is a second reason founders underestimate the problem. Lemkin has pointed out that early-stage companies often win nearly every deal they are in, because they are only invited to a small share of them. Win rates falling after roughly $1 million to $2 million in ARR, he argues, is a healthy sign that the company is getting into deals it is not yet ready for.
The practical consequence is that a flattering early win rate leaves you with no habit of asking why you lose. His fix is simple and worth copying: track loss rate, and require every open opportunity to be closed out as won or lost within a set time. A stalled deal that never gets closed is a loss nobody has admitted to. If you have not yet cleaned up ghost deals, the pipeline hygiene method is the place to start, because a debrief programme built on a dirty pipeline will chase the wrong deals.
Replace the dropdown with a debrief
At 10 to 40 qualified decisions a year, you do not have a data set. You have a short list of case studies, and each one deserves a conversation. The recommendation here is one process, run by the founder:
- Make the reason pending. When a deal closes lost, set its reason to "unconfirmed". The field cannot hold a real reason until the buyer has spoken or fourteen days have passed.
- Write your own guess the same day. One line, in your own words, on why you think it was lost. You will compare it later, and the distance between your guess and the buyer's answer is the most useful number this process produces.
- Ask within 48 hours. Email whoever made the decision. If you only know your champion, ask them who decided and whether you can speak to that person. Keep the message short, and say plainly that you are not trying to reopen the deal.
- Hold a 20-minute call within two weeks. A call, not a form. A form collects the polite answer in writing.
- Run the call yourself, where you can. A founder has no quota in the conversation and nothing left to sell, which can make buyers more candid than they would be with a rep. If you were also the seller on that deal, buyers may soften their answers, so let a co-founder or advisor who was not in the deal make the call instead.
- Record their words, then classify. Quote first, label second.
- Do the same for wins. A won deal tells you which part of your story landed and what nearly broke the deal anyway.
Questions that get past the polite answer
The most common mistake is asking the question you already have an answer to. "Was it price?" invites a yes and ends the inquiry. Ask for the situation instead:
- What had to be true for you to say yes?
- What did you do instead, and what made that the easier choice?
- When did you stop expecting to choose us?
- Who else had to agree, and what did they need to see?
- If we came back in six months, what would have changed?
Then stop talking. The silence after the third question is where most of the useful answers arrive.
Why deals are really lost: sort answers by where the deal broke
Classify each answer by the point in the buyer's decision where it failed, not by the label a CRM would offer. Five categories cover most of what you will hear, and each one points somewhere different.
| Where the deal broke | What the buyer says | Where to look |
|---|---|---|
| Urgency | "It wasn't a priority this quarter" | How you establish cost of inaction and whether you qualified for a real trigger. See deal scoring |
| Confidence | "We weren't sure it would work for us" | Proof, references, pilot terms |
| Process | "Security, legal or another stakeholder stopped it" | Your readiness for procurement and committee roles |
| Fit | "It solved a problem we only half have" | ICP and qualification, not the demo |
| Price | "We wanted it, but could not justify that structure" | Packaging and tiers. See building tiers from deal history |
An illustration, clearly hypothetical: a founder records a loss as "price". On the debrief call, the buyer says the quote was acceptable, but the finance lead asked what happens if the first quarter does not work, and nobody could answer. That is a confidence problem. The right fix is pilot terms or a reference customer, and a discount would have repaired nothing.
Do not act on one call. Act on the repeated word. If the same phrase appears in three debriefs, you have a pattern worth changing something for this week. With a handful of deals each one is a case, not a data point, so look for repetition rather than averages.
Deal history is only as good as its labels. Reading your last twenty deals for pricing signals, as the pricing-tiers piece recommends, works better when the loss reasons in those deals have been confirmed by the people who made the decision.
What if they won't talk
Some buyers will not take the call, and a founder with fifteen decisions cannot claim statistical proof from the ones who do. Both objections are fair. A call may be declined, and the buyers who agree are not a random sample. Even so, a partial set of confirmed reasons is better than a full column of unconfirmed ones, because it replaces a guess with something a buyer actually said. Keep the ask short, keep it free of any sales motive, and record declined requests too, since a buyer who refuses to explain is itself a data point about the relationship.
The recommendation changes at volume. With hundreds of decisions a year, surveys and third-party interviewers become sensible, and statistical comparison starts to mean something. Below that, the founder's call is the better instrument.
Five calls before you change anything
Before the next pricing change, demo rewrite or roadmap argument, run five debriefs. Compare each buyer's answer with the one-line guess you wrote on the day. If your guesses were right, you have lost nothing but a few hours. If they were wrong, you have just saved a quarter of fixing the wrong thing.
If the same upstream failures keep appearing (urgency that was never established, a committee you did not map, a process you were not ready for), the issue sits in the system rather than in any single deal.
Want a second opinion on why your deals are really being lost?
RivoAxis works with early-stage and growth-stage founders on the sales systems behind patterns like this one.
Talk to RivoAxis