Blast your customer list with an offer and some of them will buy. That's not in dispute and it never was.
A few take it. Somebody screenshots the number. And the conclusion writes itself: see, it works.
What that sentence is actually doing
"It worked" is the most durable defense of a bad method in business, because it's always true and it never counts the other side.
The few who bought are visible, attributable and countable. They show up in a report with names on them.
What doesn't show up: the customer who read the offer while three tickets were open and decided you don't know what's happening in their account. The one who was two months from being genuinely ready and now associates the whole category with being sold at. The dozen who didn't get angry, just quietly moved you to the pile of senders they skim past, including on the messages that were trying to help them.
None of those become a line item. There's no field for it, no meeting where it comes up, and no way to attribute the renewal you lose eight months later back to the email that started it.
So the ledger has revenue on one side and nothing on the other, and it balances every time.
Then the second thing happens
The results are underwhelming, because a handful of sales against a whole list is a rounding error against the work.
And the lesson somebody takes from that is not "we did it badly." It's expansion doesn't work. Too slow, too much effort, customers don't want it, we tried.
That's the expensive part. Not the campaign, which costs a bit of goodwill and produces a bit of revenue. The conclusion. A company runs the crude version once, gets the crude result, and files the entire motion under things that don't work here. Then nobody revisits it for three years.
The good version and the bad version got the same name, so the bad one poisoned the well for both.
Why anybody does it anyway
Impatience, mostly.
The deliberate version takes longer, and "longer" gets heard as "long." It isn't. It's a few weeks of knowing which accounts are approaching a readiness gate instead of sending to everybody on Tuesday. But the blast can happen today, and today is very attractive when there's a number to hit.
So you get the fast version, then the small result, then the wrong conclusion, in that order, reliably.
The general form
This isn't really about email.
It's a pattern that protects any method with a visible upside and a diffuse cost: the benefits are attributable and the damage isn't.
Anything shaped that way survives review indefinitely, because every review looks at what can be counted. The people harmed by it are not in the room where it's evaluated, they're distributed across a customer base, and the harm surfaces later in a form that gets attributed to something else.
You can find that shape all over a business once you're looking for it. The aggressive close that hits the quarter and sets a churn date. The discount that lands the logo and permanently lowers the base. The mandatory upgrade that holds NRR flat and hands somebody a reason to leave.
Every one of them worked. That's the problem.
What to do about it instead
Ask what a method costs when it fails, not just what it produces when it succeeds.
For a blast: what does it do to the people who weren't ready? If the answer is nothing, send it. If the answer is that a chunk of your base learns you don't know where they are, you just bought a few sales with the channel you'll need next quarter.
Then run the other version once, properly, on the accounts that actually reached a gate. It's a smaller send and a much better rate, and the number it produces is the honest comparison the "it worked" defense has never had to face.
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.