A company in Brazil was showing me their dashboard. Everybody has a dashboard.
Theirs collected customer reviews for e-commerce brands. You'd pick a product and see how it was doing. On the product we were looking at, the whole panel was bright red, because the number of reviews was falling.
The product was flip-flops. We were coming out of summer.
Fewer people were buying flip-flops, so fewer people were reviewing them. Even if every single buyer had left a review, the number still falls when the sales fall. That red panel wasn't measuring anything about reviews. It was measuring the season.
Three things are wrong there and they stack.
The first is that the alarm was false. Nothing was broken.
The second is what the alarm implied. This is a tool whose entire job is generating reviews, and it had put "your reviews are falling" in red at the top of the screen. The product was announcing that it wasn't working.
You can defend that. It's honest, it shows them what's true rather than what flatters us, and I'd normally be the first person arguing for it.
The third thing is the one that matters. Whoever opened that screen was going to spend their attention on the red thing. Not because it was important. Because it was red, and it was there.
The done-for-you problem
A company I worked with ran a done-for-you outbound service. The customer paid specifically so they wouldn't have to do the work.
The platform underneath was white-labelled, which meant there was no real way to decide what a customer could see. They got everything the operator got.
So they'd log in and find open rate sitting at zero. You don't track opens in cold email, because the tracking pixel is one of the things that gets you filtered. The number reads as total failure and it means nothing at all.
Then bounce rate, which was fine, and looks alarming to anyone who hasn't spent years staring at it.
Then the replies. A reply is the whole goal, so a folder full of them looks like progress right up until you read them. Out of office. Take me off your list. Worse than that, some of them.
Why is this happening to me. That was the question, and it's a fair one.
Those hostile replies are good news, which is genuinely interesting. When you can't measure opens, a rude answer is the best proof you have that mail is landing in inboxes rather than being filtered.
That's a real insight and it took ten minutes to explain. Every time. To every customer.
The only number that mattered was positive replies. Really the only thing that mattered was the inbox where the customer worked them. They got that. They also got forty other things, and the forty other things were where the conversations went.
It got worse when results were slow, because then customers did what any reasonable person does. They went digging.
Worth being clear about what that is. It isn't suspicion. They're trying to help. They've paid for an outcome, it hasn't arrived yet, and they've got a screen full of information in front of them, so they go looking for the problem. That's the behaviour of somebody invested in this working.
Campaign detail was right there, including the scripts.
Those scripts used spintax, which is a syntax for generating variations, and it reads like garbled nonsense unless you already know what you're looking at.
So a customer already worried about results would open the thing meant to reassure them and find what looked like broken copy written by a machine.
Here's the thing about all of it: nobody chose any of it. If we'd sat down and designed a screen for that customer, almost none of this would have been on it. It was there because it came in the box, and what comes in the box is a decision somebody else made about a different user.
The law
Whatever you put in front of a customer becomes the most important thing to them.
Not the most important thing. The most important thing to them. Those two come apart constantly, and the gap between them is where all the damage lives.
You don't get to add a caveat afterwards. You don't get to say the red number isn't the real story, or that open rate doesn't apply here, or that the scripts look like that on purpose. You can say all of it and you'll still be arguing against something they're looking at.
The screen wins. It's right there and you're not.
Executives don't ask
There's a version of this that costs more, and it shows up with senior people.
Put something in front of an executive that they don't understand, and you are not going to get a question. I know that's a generalization. I'll stand behind it anyway.
They're not going to say "what am I looking at." They'll nod, thank you, and put it away. Nobody gets to that level by asking what the chart means in front of other people, and they've got twelve other things today.
So you never get the follow-up where you explain it. You get politeness, and then nothing, and you don't find out why. The whole thing reads as a success right up until it quietly isn't one.
That's the expensive version, because it produces no signal at all. A confused customer who complains is doing you a favour. A confused executive just stops.
And if the thing you handed them was meant to open a relationship, understand what actually happened. A document nobody understands isn't a conversation starter that underperformed. It's a conversation ender. It did the opposite of its job, and it did it politely, so you'll never know.
I'm doing this right now
I'm building something that reads a company's sales calls and tells them what those calls are doing to the lifetime value of the customers on the other end.
Underneath it there's scoring. Criteria, severities, distributions, decimal places. It works, and it's the reason the output is worth anything at all.
None of it should ever reach a customer.
I got the first version wrong in exactly the way this post describes. I put a mean severity on a nought-to-two scale in the same cell as a count of calls, so "0.60" sat above "4 of 10" with nothing explaining either. I know what that means. Nobody else would, and honestly I wouldn't have known what to say about it on a call.
The problem wasn't precision, which is what I assumed at first. Rounding 0.60 to "about half" would have fixed nothing. The problem was two incompatible quantities stacked as though they were one, with the reader left to work out which was which. That's a more useful thing to know, because the fix for over-precision is rounding and the fix for this is deciding that one of the two numbers shouldn't be there at all.
That's the tell. If the person who built the thing can't defend a number out loud, it isn't ready to be seen.
So what goes in front of somebody is a page. What's already working, what's capping it, and what to do instead sitting next to each one. Counts rather than scores, because "in 17 of 19 conversations" needs no explanation and 1.47 needs a paragraph.
The scoring didn't go away. It moved behind the wall, which is where it was always supposed to be.
Short costs more
Now the part nobody wants to hear.
Dumping the raw thing is easy. The scores, the distributions, the decimal places, the full analysis. All of it is already computed, so putting it in front of someone costs nothing and it looks generous.
Turning it into something a person reads once and acts on is genuinely hard, and it's hard in a way that doesn't look hard afterwards. That's what makes it easy to skip.
If I'd had more time I'd have written a shorter letter. That gets attributed to Mark Twain and probably wasn't him, and it's still true. Short costs more.
There's real power in the raw version. That's not the argument. The argument is that power in a form nobody can use isn't doing anything, and power in a form somebody misreads is doing damage.
The test
It isn't whether the thing is true. The flip-flop panel was true.
It isn't whether you have it. That company had every number a reviews tool could produce, and the customer got all of them.
It isn't whether showing it is honest. Honesty is the argument people reach for when they haven't thought about what happens next.
The test is whether they should care.
If they shouldn't, don't put it in front of them.
Lincoln Murphy formally named and popularized Customer Success starting in 2010 and has spent 15 years connecting it to expansion revenue and commercial outcomes. Read The Premise.