Win Loss Analysis: Interview the Decision, Not the Deal

Ask a founder why they lost their last big deal and the answer comes back in under two seconds. Price. Missing integration. Bad timing. Went with the incumbent.

Now ask the buyer. You will usually get a different answer.

One win loss vendor sampled around 3,000 lost deals from a database of more than 50,000 buyer interviews and compared the reason the rep had typed into the CRM against the reason the buyer gave in the interview. The two agreed roughly 15 percent of the time.

That gap is the whole reason win loss analysis exists. It is also, in most companies, the reason it does not work: the programme gets built to collect reasons, when the useful thing was never the reason.

01 What Win Loss Analysis Actually Is

Win loss analysis is the practice of interviewing the people who bought and the people who did not, then using what they say to change your positioning, pricing, sales process and product.

Read that again for what it does not say. It is not a report on lost deals. It is primary research into how one specific market makes one specific decision, and lost deals are only half the sample.

Here is what most teams do instead. Someone pulls the closed-lost report, groups it by the loss reason field, sorts descending, and takes the tallest bar to the leadership meeting. Price is the tallest bar, because price is always the tallest bar. A discount policy gets loosened, two competitor features go on the roadmap, and nine months later the win rate has not moved.

The dashboard cannot tell you three things, and they are the three that matter.

Which alternatives you were actually in the room with. Not the two competitors in your battlecards. The real set usually includes an internal build, a spreadsheet somebody already maintains, and doing nothing for another year.

Which claim did the persuading. Every deal has one sentence that moved the buyer from interested to committed. It is rarely the sentence at the top of your homepage.

When the decision was really made. Your CRM says the deal closed in stage five. The buyer will tell you they had privately picked a favourite in week two, and the rest was procurement.

None of those live in a field. They live in a conversation.

02 Your CRM Loss Reasons Are Guesses

It is worth being precise about why the CRM data is bad, because the usual explanation is that reps are lazy, and that explanation leads to the wrong fix.

The rep is not lazy. The rep is guessing, and they are guessing for structural reasons.

The buyer did not tell them. People who have chosen someone else do not sit down and explain why to the vendor they rejected. They go quiet, or they say the kind thing. “You came in high” is the most graceful exit sentence in business, because it implies the product was good and the decision was arithmetic.

The dropdown was written by someone with a theory. A fixed list of eight options cannot record a reason nobody had thought of yet, and the reason you most need is exactly the one nobody had thought of yet.

One option is safer for the rep than the others. “Lost on price” is the only entry on that list that does not implicate the person filling it in. Not dishonesty. Gravity.

So the field records what a person under mild social pressure could type in four seconds. Adding more dropdown options does not fix any of that. Neither does making the field required.

There is a quick diagnostic here. If more than about a third of your losses are logged as price, you do not have a pricing problem. You have a value gap that nobody has been able to name, and price is what it looks like from the outside.

03 Interview the Decision, Not the Deal

This is the part that separates a win loss programme that changes something from one that produces a quarterly deck.

The instinct is to call the buyer and ask why they chose the other vendor. Do not open there. That question asks a person to justify a decision they have already made, to the party who lost it. What comes back is a reconstruction: tidy, rational, and assembled after the fact to sound sensible.

Walk the decision forward instead, in the order it actually happened.

One. What changed? “Take me back to the week you decided to go looking. What had just happened?” Nobody starts evaluating software on a calm Tuesday. Something broke, someone joined, a number got reported, a contract came up. That trigger is the most commercially valuable sentence in the whole interview, because it tells your demand generation who to talk to and when.

Two. Who made the list, and how? “Who else did you look at, and how did you first come across them?” This gives you the true competitive set and the true discovery channel in one answer. Push until they include the unglamorous options: the internal build, the existing tool stretched one more year, the decision to do nothing.

Three. What made one of them the safe choice? “When did you know it was going to be them? Who else had to agree?” Buying decisions in B2B are rarely won on preference. They are won when one option becomes the one nobody can get blamed for. Find the moment that happened and find the person who had to be convinced.

Only then, at the end, ask about you. “What would have had to be true for us to be the obvious answer?” By that point they have talked themselves back into the actual sequence of events, and the answer is worth something.

The order is the mechanism. The early questions ask for memory. The late questions ask for opinion. Ask for opinion first and every memory that follows gets quietly bent to support it.

Four rules keep the interviews honest:

• Never the rep who owned the deal. Buyers will not say the awkward thing to the person they turned down, and the rep cannot hear it neutrally. A founder, a product marketer or an outsider all work.
• Thirty minutes, recorded, transcribed. You are collecting language, not conclusions.
• Write down the exact words. The phrases buyers use are the raw material for your headlines. Do not translate them into your own vocabulary on the way into the notes.
• Say what it is when you ask. Research, not a save attempt. Response rates are better when the buyer believes it, and it is only believable if it is true.

04 Call Your Wins, and Split Out the No Decisions

Two segmentation mistakes quietly ruin most of these programmes.

The first is only calling losses. It feels efficient, since losses are where the problem is. But losses only tell you where you slipped. Wins tell you what worked, and specifically they tell you the sentence a buyer used to justify the spend to somebody who was not on the calls. That sentence is your landing page. You cannot get it from a lost deal.

The second is treating “lost” as one bucket. A deal lost to a competitor and a deal lost to no decision are different diseases with different cures. Losing to a rival is a positioning problem: two options were on the table and yours read as the riskier one. Losing to no decision is a status quo problem: nobody was convinced that another year of the current mess was expensive. No feature has ever fixed the second one. In many B2B markets the no-decision pile is the largest single bucket, and averaging it into the competitor losses hides the biggest available gain.

Three groups, then, and they get counted separately: won, lost to a named alternative, lost to no decision.

On volume, eight to twelve interviews per group is enough to claim a pattern. You will usually hear the same sentence for the third time around interview six. And it works far better as a standing habit than a project. Four to six interviews a month, forever, beats forty interviews once a year, because the market moves and a year-old finding is a year-old opinion.

05 Turn Findings Into Four Changes, in Order

A finding is not a change. The last step, the one that gets skipped, is forcing every theme into one of four buckets, ordered by what it costs to act on.

Words. Claims, headlines, the way the category gets described, the objection handled on the call. Free, and you can change it this week.

Sales process. What gets shown and when, who is in the room by the second call, which question gets asked on discovery. Cheap, a fortnight.

Pricing and packaging. The shape of the offer, the unit it is billed on, what sits in which tier. Medium cost, one quarter, and it touches everybody.

Product. The most expensive answer in the building, and the one every roadmap discussion reaches for first.

Work top down. Most findings die in the first bucket, which is the point: the majority of losses that get blamed on missing capability turn out to be losses where the capability existed and never got connected to the buyer’s actual problem.

One gate before anything reaches the roadmap: a theme has to appear in at least three interviews across two of your three groups. Below that it is a story, and stories that get built are how roadmaps fill up with features one loud prospect wanted.

Then measure the thing you were trying to move. Not interviews completed. Win rate by segment, and win rate against the specific alternative you were losing to. If those two numbers do not move within two quarters, the programme is producing reading material.

The output of good win loss analysis is not a report. It is a shorter list of things you were wrong about, and the cheapest one to fix sitting at the top.

Want the interviews run on your last twenty deals? We tear down your GTM in five days. Book a teardown →

Frequently asked

What is win loss analysis?

Win loss analysis is the practice of interviewing the buyers who chose you and the buyers who did not, then using what they say to change positioning, pricing, sales process and product. It is primary research into how a market decides, not a report on lost deals.

How many win loss interviews do you need?

Eight to twelve per segment before you call anything a pattern. In practice the same sentence usually shows up for the third time somewhere around interview six. Running four to six a month as a standing habit beats one big annual project.

Who should conduct win loss interviews?

Anyone except the rep who owned the deal. Buyers will not say the awkward thing to the person they turned down, and the rep has an interest in the answer. A founder, a product marketer or an outside interviewer all work.

What questions should you ask in a win loss interview?

Walk the decision in the order it happened. What changed in the business that made you start looking, who else made the list and how you found them, and what made one option the safe choice. Save any question about your own product for the end.

Why are CRM closed-lost reasons unreliable?

The rep picks from a dropdown someone else wrote, using a guess, because losing buyers rarely explain themselves. One sample of about 3,000 lost deals found the recorded reason matched the buyer's stated reason roughly 15 percent of the time.

How is win loss analysis different from churn analysis?

Win loss analysis studies the buying decision, so it tells you about positioning and competitive set. Churn analysis studies the renewal decision, so it tells you about value delivered after the sale. The two share almost no root causes.