Every closed-lost record in your CRM has a reason attached to it. Someone picked that reason from a dropdown, in under ten seconds, on the same afternoon they got the bad news. In most pipelines the winner is price.
That is not a loss reason. It is the rep’s guess, filed as a fact.
Win-loss review is the cheapest research a services firm can run, and almost nobody runs it properly. The blocker is rarely budget. The work gets handed to the one person structurally incapable of doing it, which is the rep who lost the deal.
The lost-reason field collects the wrong data
HubSpot ships a Closed Lost Reason property. Salesforce has its Loss Reason picklist. Both are set by the deal owner, both default to a short list of tidy options, and neither one gets audited by anybody. The rep picks the option that implicates the rep least. Nobody has ever been put on a performance plan for losing on price.
Run the numbers on a firm working 120 qualified opportunities a year at a 22% win rate. Ninety-four losses. If 60% of those come back tagged price, 56 deals a year sit under one explanation. Real distributions are almost never that lopsided. Call it a quarter genuinely price, a quarter timing or no decision at all, and the rest split between a capability you failed to demonstrate and a decision-maker you never met. That puts roughly 40 losses a year in the wrong bucket, and it is always the bucket that asks nothing of you.
Price is a comfortable finding. A price loss says the offer was right and the buyer was cheap. No part of that story requires anyone to change how they sell next quarter.
Buyers answer when the asker has nothing at stake
The buyer who told your rep no will frequently explain the decision to somebody else at your firm. Not to the rep. Turning the same person down twice inside a month costs the buyer something socially, while answering a neutral third party costs them twenty minutes and nothing else.
So take the interview away from the deal owner. Give it to whoever runs revenue operations, or the founder, or a rep who covers a different segment and has no reason to defend the calls that were made. The script matters less than the distance.
Timing matters more than most firms assume. Ask inside ten business days of the decision and the buyer still remembers the specific moment they went cold. Ask at 60 days and you get a rationalization, tidied and made coherent by hindsight. The detail worth having is the one already fading.
Send the request from the neutral party, name the deal, ask for twenty minutes, and say plainly that you are not reopening anything. Expect somewhere between a third and half of recent losses to take that call. A written reply from the rest still counts as evidence.
Four questions do most of the work
Skip the survey. A five-point scale on “value for money” produces a number you can chart and cannot act on. Ask these instead, in this order, and let the buyer run long.
- What was going on in your business when you first took our call? The answer is the real trigger event, and it belongs in your targeting.
- Who else was in the room when the decision got made, and what did they push for? You are listening for the name you never met.
- What did we say that made you doubt us? Buyers answer this one with uncomfortable honesty, and it is the only question that reaches the pitch itself.
- What would we have had to show you to win? The verb is show. Buyers almost always name a specific artifact, and usually one you could have built in an afternoon.
Never ask whether price was the issue. Hand a buyer a socially acceptable exit and they will take it, and you will have spent twenty minutes confirming the picklist entry you already distrusted.
Write the notes as quotes, not as summary. “We could not tell who would actually be doing the work” is usable on Monday. “Concerns around delivery model” is what that same sentence becomes once somebody helpfully cleans it up, and it points at nothing anyone can fix. Keep the buyer’s language intact. One page per interview, quotes at the top, your interpretation underneath and clearly labeled as yours.
Sampling only your losses teaches half the lesson
Run the same four questions on won deals. Firms skip that half because a win feels self-explanatory. It never is. The reason you won is often something nobody on your side chose on purpose: a case study that happened to match their sector, one reference call, the fact that somebody answered the phone on a Sunday.
Wins also correct a bias baked into loss-only research. Interview twenty losses and you build a catalog of everything that repels a buyer, then quietly reshape the pitch around objections raised by people who were never going to buy. Keep the ratio near two losses to one win and the picture stays honest.
The output is a change to the pipeline
A win-loss program that ends in a slide deck is a hobby. Every finding routes to one of four destinations, and a finding that routes nowhere gets dropped.
Targeting. Trigger events from question one go straight into the ICP definition and the list build. When eight of twelve buyers describe the same event, that event is a filter, not an anecdote.
Stage criteria. A committee member you never met is a qualification failure, and the fix lives in the exit criteria your stages are supposed to enforce. A stage that lets a deal advance without a named economic buyer will keep producing that same loss on a schedule.
Materials. When four buyers name the same missing artifact, build it once and give it to everyone. Do that for a year and the proposal becomes a formality, because the persuasion already happened.
Qualification. Losses that trace back to a mismatch visible in week one are a direct argument to disqualify faster. Same for deals that died in silence rather than in a decision, which almost always means the follow-up had no structure holding it together.
Fifteen interviews a year at twenty-five minutes each is roughly six hours of one person’s calendar, plus the writeups. Set that against a pipeline where one misdiagnosed pattern quietly costs two or three deals a quarter. The trade is not close.
There is a version of this that goes further, where the calls get recorded in Gong and the language buyers use gets fed back into the outbound copy. Worth doing eventually. Not worth waiting for.
Here is the check you can run this week. Pull your last twenty closed-lost records and read nothing but the reason field. Count how many contain an instruction, meaning something a person who never touched the deal could act on tomorrow morning. If fewer than five clear that bar, the field is decoration and the research has not happened yet.
Florida Man Innovations builds and grows revenue systems for firms that want a pipeline that works without heroics. If your loss data cannot tell you what to change, that is where the work starts. More at brettfl.com, calendar at meet.brettfl.com, or write to [email protected].
Ask the buyer.