Why did our win rate drop? What to check when every CRM slice comes back flat

The loss review is forty minutes in and every chart is flat.

Win rate is down nine points since spring, so we slice it. By segment, by lead source, by rep, by deal size, by the loss reason field. Each cut shows the same thing, a drop spread evenly across everything, as if the whole pipeline got slightly worse at once.

A flat slice is information. It usually means the cause lives somewhere the CRM has no field for, and the first place we'd look is what buyers were comparing us to.

The market can change what we get lined up against while our positioning stays exactly where it was. No dropdown records that. The way to find it is to read ten losses and ten wins and write down, for each one, what the buyer thought the alternatives were.

It takes one person about two days.

What does the CRM actually record?

The deal. Stage dates, amount, source, owner, a competitor if the rep picked one, and a loss reason chosen from a list after the call that ended it.

Slicing those fields is the right first move, and it often works. If the drop sits in one segment, one rep cohort, or one deal-size band, the cause is in the deal and the slices will find it.

New reps are still ramping. A channel started sending worse leads. We moved upmarket before the product was ready for it. If one of those cuts lights up, stop reading and go fix that.

But look closely at the loss reason field when nothing lights up. A rep fills it in, from a dropdown somebody built two years ago, about a decision made in a meeting the rep didn't attend. "Price." "Timing." "Went with competitor."

The rep is being honest. They're reporting the last thing they heard, and the last thing a buyer says is a polite summary. The reasoning stayed in the room.

A grocery receipt lists everything that went in the bag. It can't show what we picked up, read the back of, and put back on the shelf.

The CRM is the receipt, and the decision happened at the shelf.

The deal changed, or the set changed

There are two kinds of win rate drop, and they look identical on a dashboard.

When the deal changed, something on our side of the transaction moved. The price went up, a strong rep left, a competitor shipped the feature we used to win on. The buyer is still making the same choice buyers made last year, and we've become a weaker option inside it.

When the set changed, buyers are making a different choice. Same product, same pitch, same price, placed next to different things.

The first kind shows up in CRM fields, because the fields describe our side. The second kind can't, because it lives in the buyer's head and it was formed before we had a record to write in.

6sense's 2025 Buyer Experience Report found that 95 percent of the time, the winning vendor is already on the Day One shortlist. The comparison has been running for a while by the time an opportunity gets created. We log the ending of a story whose first chapters we never saw.

So the two drops need opposite work. One sends us into the deal data, and the other sends us out to the buyer.

Comparison-set drift

Comparison-set drift is what happens when the things buyers line us up against change and our positioning doesn't.

Here's an invented example, though the outline will be familiar. Picture a customer support platform at $8M ARR. Two years ago, nearly every deal was a bake-off against two other help desks, and the positioning was built for that fight. Faster setup, better reporting, friendlier pricing.

It won a lot.

Then three things happen that have nothing to do with the company. The CRM vendor their buyers already pay bundles a support module into the contract. An AI agent startup starts pitching the same VP on needing far fewer seats. And a few buyers begin to wonder whether a shared inbox and a chatbot would be enough.

The homepage still says faster setup and better reporting. Both claims are still true. They answer a question that fewer buyers are asking.

Nothing in the CRM moves. The competitor field shows the same two names, or it's blank, because a bundled module and "do nothing" were never in the picklist. The loss reason says price, because next to something the buyer already pays for, everything is expensive.

That's the same mechanism behind the "too expensive" objection, and it's why so many losses that look competitive are really deals that went to nothing.

Drift has no launch date. It arrives a deal at a time, in every segment and for every rep, which is exactly why the slices come back flat. A cause that touches everything a little can't be isolated by cutting the data into parts.

It also can't be fixed from inside the old frame. Competitive positioning is a claim about where we stand relative to something. When the something changes, the claim stops being an answer, however well it's written.

Ten losses, ten wins

Pull your ten most recent closed-lost deals and your ten most recent closed-won, limited to ones with recorded calls or a buyer who'll still pick up the phone. Take them in order. Choosing the interesting ones is how we end up confirming what we already believed.

For each deal, write down four things, in the buyer's words wherever you can:

• What they were comparing you to, including things that aren't vendors (the tool they already own, a hire, a spreadsheet, waiting a year)
• The phrase they used to describe what you are
• When the comparison first showed up (the first call, a late email, a question from finance)
• How it ended

Start with the first discovery call. Buyers tend to say the set out loud early, in lines like "we're also looking at" or "we already have something that sort of does this." A prospect who calls you just like a competitor, but worse has handed over the comparison set for free.

After the recordings, read the email thread. Then, where the record is thin, ask for fifteen minutes with the buyer.

Two questions are usually enough. "When you started looking, what were the options on the table?" And then, "What did you end up doing?" It's worth resisting the urge to ask why you lost. That question gets the dropdown answer again, delivered more kindly.

A fuller win-loss program asks much more, and this read is deliberately smaller than that.

The wins matter as much as the losses. A loss tells you a set exists. Only a win tells you which set the current positioning still works in, and without that, there's nothing to measure the losses against.

One person should read all twenty, ideally someone who wasn't carrying quota on any of them. Twenty deals split across four readers gives you four vocabularies and no pattern.

How do you read the result?

Put the twenty rows on one page and sort them by comparison set. Then look at the outcome column.

If the wins cluster in one set and the losses sit in others, that's drift. The positioning still wins the fight it was built for, and that fight is happening less often.

More selling won't close that gap. What's left is deciding which of the new comparisons to answer directly. That's a positioning question, and it's close to the one in product problem or positioning problem.

If wins and losses share the same set, the buyers are all making the same choice and landing on different sides of it. Then the cause is in the deal after all. Pricing, execution, or fit.

Go back to the slices with a sharper question, and look hard at who the losing accounts were, because a profile can match on paper while nothing inside the account was forcing a decision.

Sometimes the page is a mess. Twenty deals, nine different comparisons, no clustering at all. That's a finding too. It can mean buyers don't know what you are, so each one files you somewhere different.

Twenty rows won't prove anything. They're a reading, good enough to say where to dig next and no better. If the page points clearly one way, it might be worth pulling twenty more before anyone rewrites a homepage on the strength of it.

What to hold until the read is done

The hard part of a falling win rate is the pressure to do something by Friday.

The usual somethings are a discount, new battlecards, an objection-handling session, and a homepage refresh. Every one of them is a fix for a cause on the deal side. If the set moved, the discount agrees with the new comparison, the battlecards arm reps against a rival who's showing up less, and the homepage gets a fresh coat of the old answer.

So it might be worth a two-week hold on all four.

Name the reader on Monday. Pull the twenty deals the same day, block two days for the recordings, and book the buyer calls for the following week. Bring one page to the next review.

If the slices already found the cause, none of this applies. Go fix what they found.

But if every chart is flat, stop cutting the receipt into smaller pieces. Go stand at the shelf.

What to Do Next

If your win rate fell and the team has been slicing CRM fields for a pattern that won't appear, you're probably looking for the cause in the one system that can't hold it. Nothing in a dropdown says what the buyer thought you were, or what they set you beside.

Run the read first. If it shows your losses being compared to different things than your wins, a Bare Strategy positioning audit picks up from that page and works out which comparison to answer and what the positioning has to say to win it. It starts with a discovery call.

If that's where you are, start here. The first conversation is free.