How to Know If Your Messaging Is Actually Working

We shipped the rewrite six weeks ago. New homepage, new one-liner, a deck the whole team agreed was better. Since then, nobody has told us it's wrong.

That isn't a passing grade. Nobody complains about a message they didn't understand. They close the tab and go read somebody else's site.

A rewrite is working when your language starts coming back to you from places you didn't put it. A stranger's first inbound email uses your phrase. Then behavior shifts, the wrong-fit demos thin out, the second call books faster. Then, a full sales cycle or two later, the revenue numbers move. Those three signals arrive in that order and at wildly different speeds, and reading them out of order is how good messaging gets killed at week four while mediocre messaging survives a year.

No complaints is not a data point

Silence used to carry information. When most of an evaluation happened in conversation, a confused buyer had to ask a question in order to keep going. The confusion surfaced because the process forced it to surface.

That structure is gone. Gartner's March 2026 survey of 646 B2B buyers found that 67% prefer a rep-free experience, and that 45% used AI somewhere in a recent purchase. Two-thirds of the people sizing us up are deliberately routing around the one channel where their confusion would have become audible.

So the confused buyer never asks. They leave.

And from our side of the glass, a buyer who left quietly looks identical to a buyer who understood us perfectly. Same absence. Same quiet inbox.

The received wisdom has a real basis and it's worth conceding. If a message were catastrophically broken, somebody would have heard by now. Sales would be loud. That holds for catastrophe. It doesn't hold for the ordinary case, which is a message that's mostly clear, lands for the people already inclined to like us, and reads as interchangeable to everyone else. That version generates no complaints at all. It generates a slightly worse quarter, which we then attribute to the market.

We read silence as consent because the alternative is expensive. Checking properly means admitting that eight weeks of work might not have moved anything.

Why can't you just ask buyers if your messaging landed?

You can. The answer will be confidently wrong in a predictable direction.

People are good at telling us what they think of words and bad at telling us what those words did to them. Show someone two headlines and they'll pick one and give a reason. That's a preference, and preferences say very little about what a person with eleven seconds and three other tabs open will actually do.

Before launch that's still useful, because we're comparing options against each other and comparison is something people can genuinely do. Validating a message before it ships is a real discipline with its own methods.

After launch the question changes shape. It becomes whether the thing already shipped moved anyone, and nobody answers that by introspection, buyers included.

There's a second problem underneath. 6sense's 2025 buyer research reported that 85% of buyers had previous experience with the vendor they bought from, and 85% of those said that experience took the form of a previous evaluation. Your buyer formed a read on you before the rewrite existed, sometimes years before. Ask them what they think of the new message and many of them will hand back a polished version of what they thought of the old one.

The echo test

A message is working when it comes back.

That's the whole test. Approval counts for nothing here. What counts is evidence that somebody other than us can reproduce the claim with nobody from our side in the room. Language we have to keep supplying isn't positioning, it's a script we're reading aloud.

Call it the echo test. Stop asking whether the words are good and start listening for them returning from somewhere you didn't send them.

They come back in three forms at three speeds. Words come back in days. Behavior changes over weeks. Revenue moves after a sales cycle or two. Most of us check the third one first, find nothing, and conclude the rewrite failed. It was never going to be there yet.

The words come back first

Say you're an observability company and the rewrite installed one claim, that on-call engineers are drowning in alerts nobody reads. The language echo is the inbound email three weeks later, from someone you've never spoken to, describing their own week in almost those words.

Listen for four things.

A prospect using your phrase in their first message, before any call. Your own reps using it unprompted, which is a harder and more honest test than whether they liked the training. The phrase turning up somewhere you don't control, in a review, a forwarded Slack thread, a job posting. And a champion explaining you to a colleague in your framing rather than inventing their own.

That last one is the highest-value signal in the tier and the one almost nobody instruments. If a buyer has to build their own explanation of us before they can carry us into a committee, the message reached the person on the call and stopped there. That gap is why positioning stops showing up on sales calls even when the messaging doc is excellent.

A distorted echo is information too. If the market keeps repeating your third bullet instead of your headline, it has told you which claim it actually wanted. Take the note. Buyer language usually beats ours anyway, which is the whole argument for voice of customer research as an input rather than a validation step.

The behavior changes next

A rewrite that works will often make the numbers look worse for a month.

Sharper messaging repels more people, and it's supposed to. If new positioning names a specific buyer and a specific fight, a chunk of old inbound now correctly identifies itself as the wrong fit and doesn't book. Demo volume drops. Somebody in a Monday meeting calls that a problem, and we usually agree with them too quickly.

Four behavior signals worth pulling.

Who books, measured as fit rather than count. Time from first call to second call, which tends to compress when a buyer arrives already convinced the problem is real. Where the drop-off sits, because clear messaging rarely removes drop-off and mostly moves it earlier. Losing people on the homepage instead of in week five of an evaluation is a large financial win wearing the costume of a scary funnel chart. And how many discovery calls open with the buyer framing the problem the way you frame it, before your rep says anything.

The demos a good rewrite costs you are the ones the team was going to lose in week six anyway, more slowly and at greater expense.

The money moves last

Win rate, sales cycle length, and average deal size are the honest tier. They're also the tier nobody can read for at least one full sales cycle, and probably two.

The reason is mechanical. Every deal in the pipeline today was sourced under whatever we were saying when it entered. Mixing pre-launch and post-launch deals into one win-rate number produces an average of two different messages and describes neither. Cohort by the date the opportunity was created, and refuse to compare across that line until the older cohort has closed out.

Gartner's same 2026 research found that confident buyers are twice as likely to report a high-quality deal as buyers with low decision confidence. That's the mechanism this tier measures. Clarity works by letting a buyer feel certain enough to sign, and that certainty surfaces later as a shorter cycle and a cleaner close.

Once the cohort is old enough to read, the sharpest instrument is win/loss interviews against that cohort specifically. Ask the buyer to describe what you do, in their words, and then stop talking.

How do you run an echo audit in two weeks?

Five moves, and maybe four hours of real work.

Write down the one sentence the rewrite was supposed to install in the market. If it won't compress to a single sentence, there's a writing problem sitting upstream of the measurement problem, and no audit will find it for you. A messaging house helps here, mostly by forcing the choice.

Pull ten recorded calls from the last two weeks. Count buyer utterances of the claim in their own language. Count rep utterances separately. Both numbers matter and they mean different things.

Read the last twenty inbound emails and highlight every phrase you wrote. If nothing is highlighted after a month of real traffic, that's a genuine signal and it arrived early enough to act on.

Cohort the pipeline by creation date, before and after launch, and label the line clearly enough that nobody averages across it in a board deck.

Then write down what you expect to be true in ninety days, and date the file. The most useful part of this may turn out to be having a record of the prediction, because memory quietly reorganizes itself around whatever the quarter ends up doing.

None of it is rigorous. It might be worth two weeks anyway, given that the method currently in use is waiting to see whether anyone complains.

Six weeks of quiet tells us nothing about the message.

It tells us about our instruments.

What to Do Next

If you've shipped a messaging change and you've been reading the absence of pushback as a green light, you're in the position this post describes. There's no bad news and no confirmation either, and the next rewrite is getting scoped on a hunch.

A Bare Strategy positioning audit is the outside read on that. We go through recorded sales calls, inbound language, and cohorted pipeline data to establish whether the last rewrite is registering anywhere real, before anyone spends another eight weeks writing a new one.

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