Atiba de Souza.Start a conversation
WritingAI

The Top Tier Nobody Buys Is Still Selling for You

By Atiba de Souza

An editorial cover image for a piece on The Top Tier Nobody Buys Is Still Selling for You. The cover reads: AI The Top Tier Nobody Buys Is Still Selling for You

I hired an agency to run Amazon ads for my book

I hired a well-reviewed, respected agency to run Amazon ads for my book. I did not tell them who I was, or that I do this for a living. I wanted to be the customer.

Two weeks in, they sent a glowing report. Units sold. Movement. Momentum. It did not cover ad spend. It did not cover ACOS, the metric that should sit between 20 and 30 percent. Mine was at 1300 percent.

1300 percent ACOS against a 20 to 30 percent target. The report was not false. It was true and partial, and partial was enough to keep a buyer paying.

Here is what I want you to hear, and it is not "agencies lie." A buyer who did not know the metric would have read the glowing report, believed it, and kept paying. The report did not need to be false. It only needed to be partial.

The part that turned for me was not their behavior. It was mine. I have built a career on marketing, and I still sat in the chair and let someone else decide which number I would be shown. Nobody talked me out of my money. I never named the metric.

So before you hire a marketer, know the one metric that decides success, and ask for it by name. In writing. Before the invoice.

That is the buyer's move. Now turn it around. When you sell, you are the one choosing which numbers the buyer sees. And the number that decides everything is not the discount on the page. It is the number the buyer carries away.

The framework is The Selling Principles, and Principle #1 is running this

The Selling Principles are the universal principles of selling anything to anyone: the human reasons a persuasive pitch works, independent of product or industry. They are the method, not the content. Their purpose is in the name of the work: earning the sale by serving the buyer's real decision.

Principle #1: State precedes message. You cannot move a mind whose state is closed. Attention and openness are physiological. The message arrives second, and it only lands if the state lets it in.

A deal page is a state machine. It narrows attention. It triggers comparison. It invites arithmetic. A buyer in that state is not weighing your offer against their problem. They are weighing your number against the number printed beside it.

That is the state a fake markdown is built to produce. And once the buyer is in it, the only message that lands is arithmetic. A fake markdown serves the seller's fear of losing the sale. It does not serve the buyer's real decision. The Selling Principles say earn the sale by serving that decision, not by managing the buyer's state until they forget what they came for.

The one word this whole thing turns on

Reference price: the number a buyer carries away from your last price and measures your next one against. It is not on your price list. It is in their memory. You do not set it with a spreadsheet. You set it with what you actually charged.

Every seller in this argument holds the same asset and the same liability. Every crossed-out price either earns the next one or spends it, and the buyer keeps the balance.

The analogy that makes it click: the top tier nobody orders

Look at any price list that works. The most expensive item on it is doing its job even when nobody orders it, because it sets what every other price gets measured against. People do not judge a price on its own. They judge it against the prices sitting next to it.

That is why the top tier exists. It is the reference price. It is doing the selling for the tier below it.

But the tool has a break condition, and the break condition is where the fake discount dies. The anchor only works if the buyer believes the top item is genuinely worth what it says. Price against a number they do not believe, and nothing happens.

A crossed-out price that was never charged is the top tier nobody believes. It is an anchor made of air. It does not hold the buyer's judgment anywhere. It just teaches them that your numbers move.

This is not a marketplace problem. It is a pricing habit, and it runs at every scale, including yours.

The whole thing at a glance

Yes

No

You show a price

Buyer enters a state before your message lands

Comparison and arithmetic, not the buyer's problem

Is the reference price real?

Buyer measures against a number they believe

Next price is easier to hold

Buyer measures your number against the number beside it

Next price needs a bigger strike-through

The massage-chair sellers at a mall in Manila

At a stand in a mall in Manila, they give you a free fifteen minute massage. It is a sales tactic, and a good one. The product costs 750 pesos, about fifteen dollars. If you buy it, you get free massages for six months.

Half the price of one massage at home, for six months of unlimited. I bought it. Of course I bought it.

With six months nearly gone, I had used it twice. I was still paying 1,500 pesos for a massage at home every week.

750 pesos, about $15, for six months of unlimited. Used twice. Still paying 1,500 pesos a week at home. I did not get scammed. I got exactly what was described. What I bought was not massages. I bought the feeling of unlimited.

Here is the turn. Sold as a fixed number of massages, the same offer would have looked like a small discount, and I would probably have walked past it. Same product. Same price. One packaging decision. Unlimited served a real human need: no ceiling, no counting, no deciding whether today is the day I use it.

So when you are packaging a service, try offering it as unlimited. People value having as much as they want far more than they actually use. That gap is the offer.

Inconsistent vs consistent

InconsistentConsistent
The crossed-out price was never chargedThe anchor price is real and some people pay it
Buyer learns to wait for the markdownBuyer learns the price is the price
Every future price needs a taller strike-throughEvery future price is measured against a real one
You sold a transactionYou sold a reference point
Works onceCompounds

Rank the parts

Not every idea in this piece is equal weight, so let me rank my own.

The minor idea is that a big retailer ran a sale that was not a sale. That is a story about somebody else. The major idea is that you set a reference price every day, whether or not you are running a promotion. That is a story about you.

Now the part I am not going to soften. You are going to read this and think, "great, but I still need the sale this month." You are right, and that is the cost. Holding a real price means being willing to lose the buyer who only closes on a number you do not believe in. That buyer is real. They are in your pipeline right now. Losing them will feel like a mistake while it is happening.

There is a second cost, and it is the one nobody names. Discounting lets you skip the harder question. If the thing is not selling, a discount answers that in an afternoon. A positioning problem takes a quarter. So the discount does something for you that has nothing to do with the customer: it protects you from finding out whether the thing is worth what you want for it.

That is the rewiring. Stop treating price as the answer to a positioning question.

When not to run this play

  • You do not have a top tier yet. Do not invent one. An anchor nobody believes anchors nothing, and a manufactured premium tier is the same lie with a bigger number.
  • The price is genuinely wrong. Then fix the price. A correction is not an illusion. Treating every discount as a lie will make you stupid about pricing, and that costs more than the discount ever did.
  • Your cost to serve usage is high. Unlimited works because people value having as much as they want far more than they use. If every use costs you real money, you are funding that gap. Sell a generous number instead.
  • You are about to fake a strike-through while reading this. Do not.

Two names before Monday

This ends in work, not agreement.

As the seller. Name the buyer you expect to lose by holding a real price. Write down the person or the segment. Then name the positioning question the discount would let you avoid. If you cannot name either, you are not holding a price. You are hiding from one. The decision is not "stop discounting." The decision is "know what the discount is buying, and whether it is worth the reference price it costs."

As the buyer. Name the one metric that decides whether the thing you are buying worked. Ask for it by name, in writing, before you sign. Then decide the move if it is withheld. Walk, or set a date to walk. A seller who will not name the line has told you where the line is.

Nobody is auditing your markdowns. Your customer is. And the number they carry away is the only price that matters next time.

Questions

What people ask next

If I've already trained my customers to wait for the markdown, how do I walk that back without losing them entirely?
It depends on how willing you are to pay the near-term cost: you will lose the buyer who only closes on a number they don't believe is real, and that loss happens now while the benefit compounds later. The way through is to stop taller strike-throughs and start charging the price you mean, consistently, so the next markdown (if any) is measured against something real. Some of the trained buyers will walk; that is the price of resetting what number they carry away from you.
How do I tell the difference between a real positioning problem and a price that's just genuinely too high, before I default to assuming it's always the former?
Ask whether a discount would answer the question in an afternoon or whether it's being used to avoid a harder question that would take a quarter to answer. If the number itself is one you don't actually believe in and wouldn't want someone to pay full price for, the price may genuinely be wrong, and fixing it is a correction, not an illusion. But if you're reaching for a discount because you're afraid to find out whether the thing is worth what you're asking, that's positioning dressed up as a pricing problem.
If I don't have a believable top tier yet, what's the actual first step to building one without it looking manufactured?
Don't invent a number and put it at the top; that's a manufactured premium tier, which is the same lie with a bigger price on it. The first step is to make sure the top item is genuinely worth what it says, meaning it's priced at something real that you'd be comfortable having some buyers actually pay. The anchor only works once it's believed, so build the belief first and let the tier follow, not the other way around.
In the moment where I need the sale this month, what does holding the real price actually look like in practice, not just in principle?
It looks like watching a specific buyer walk away because they only wanted to close on a number you don't believe in, and letting them go instead of inventing a strike-through to catch them. It will feel like a mistake while it's happening, because the lost sale is immediate and the reference price you're protecting only pays off on the next deal and the one after that. In practice, holding the price means refusing to answer a positioning question with an afternoon discount, even when the discount is sitting right there as the easier move.
Outside of a formal metric like ACOS, how do I even figure out what the equivalent 'one number that decides everything' is for my own business?
It depends on what number, if left out of your own report, would turn a true statement into a misleading one. In the agency case that number was spend relative to return; in pricing it's the number the buyer carries away and measures your next offer against. Find the number that connects what you're spending or promising to what the buyer actually measures you by, name it in writing before you commit to anything, and treat any report that omits it as partial, not complete.