The headline said profit tripled
Amazon's Q1 earnings landed last week. Net income: $62.6 billion, up from $21.4 billion a year earlier. The headlines wrote themselves. "Amazon profit nearly triples." "Amazon posts record quarterly profit." The market moved. Analysts updated their models. The story was told.
Except five-sixths of that profit — $53.4 billion of the $62.6 billion total — is a paper revaluation of Amazon's stake in Anthropic. Amazon hasn't sold those shares. No cash changed hands. The value went up on the balance sheet, and accounting rules require that increase to flow through the income statement. It's real in the accounts. It's not in the till.
Operating income, the measure of what the business actually earned from trading, was $27.5 billion. Still impressive. Still up year-on-year. But not a tripling.
The distinction matters, but almost nobody made it. The first number set the frame. Everything that followed was interpreted through that anchor.
The first number always wins
Dan Ariely's anchoring experiments are canonical now, but they bear repeating because the effect is so reliable. Show someone a random number — the last two digits of their social security number, say — then ask them to estimate the price of a bottle of wine or the value of a computer keyboard. The random number pulls their estimate toward it. The higher the anchor, the higher the guess. The lower the anchor, the lower the guess.
The anchor doesn't need to be relevant. It doesn't need to be justified. It just needs to be first.
Amazon's $62.6 billion did exactly that. It arrived first. It was big. It shaped every judgement that followed, even among people who should know better. Analysts who understand GAAP, who can read a cash flow statement in their sleep, still led with the net income figure because that's what the headline demanded. The anchor had already done its work.
The lesson isn't that Amazon misled anyone. They didn't. The accounts are accurate. The disclosure is clear. But clarity doesn't override sequence. The number that arrives first wins the frame, whether or not it's the number that matters most.
Your listing runs the same machinery
Now turn that lens on your product detail page.
The strikethrough price is an anchor. It doesn't matter whether anyone ever paid it. It doesn't matter whether it was a real price or a nominal list price that existed for compliance purposes. What matters is that it's first, it's bigger, and it sets the reference point against which the actual price is judged.
The customer doesn't think, "Is this discount real?" They think, "$79.99 down from $149.99 — that's a good deal." The comparison happens automatically. The anchor does the work before conscious reasoning gets a vote.
The star rating works the same way. 4.7 stars is the first signal of quality most shoppers see. It arrives before they read a word of the bullet points. It shapes their interpretation of everything else on the page. A feature that might seem trivial in isolation becomes evidence of value once the 4.7-star anchor is set. A potential concern mentioned in a review gets downweighted because the aggregate score has already told them this is a good product.
The number that arrives first wins the frame, whether or not it's the number that matters most.
The social proof badge — "10K+ bought in past month" — is another anchor. It says: this many people chose this product. The inference is automatic: if that many people bought it, it must be worth buying. Whether those purchases were driven by a promotion, a TikTok spike, or genuine organic demand is a question most shoppers never ask. The big number landed first. It set the frame.
Even the "Deal" badge is an anchor of a kind. It doesn't specify the size of the deal or the reference price. It just signals: this is cheaper than it was. That signal is enough to pull perceived value upward. The customer fills in the rest.
Which of your numbers goes first
The anchoring effect doesn't care about your intentions. It runs automatically. The only choice you have is which number you put in front of the customer first.
If your main image is a lifestyle shot that de-emphasizes the product, you've anchored on context instead of object. If your title leads with a feature instead of a benefit, you've anchored on specification instead of outcome. If your A+ content opens with a brand story instead of a problem solved, you've anchored on history instead of relevance.
These aren't moral choices. They're structural ones. The first thing the customer sees becomes the lens through which they interpret everything else. You can't avoid setting an anchor. You can only choose which anchor you set.
So: what's the number you want to win?
If you're in a category where price is the primary decision variable, lead with proof of value — the comparison price, the per-unit cost, the lifetime saving. If you're in a category where trust is the barrier, lead with the star rating, the review count, the "Amazon's Choice" badge. If you're in a category where differentiation is hard, lead with the claim that separates you from the next twenty listings in the search results.
The mistake is letting the first number be accidental. The customer will anchor on something. The only question is whether you chose it or whether it chose itself.
The accounts were accurate; the frame still moved
Amazon's Q1 earnings report was accurate. The $62.6 billion net income figure is correct under U.S. GAAP. The footnotes explain the Anthropic revaluation. The cash flow statement shows what came in and what went out. Anyone who wanted the full picture could get it.
But the headline didn't say "Amazon posts $27.5bn operating income, up 23% year-over-year." It said profit tripled. Because the first number was $62.6 billion, and that number was big enough to be the story.
The frame moved because the anchor moved it. The same machinery that ran on analysts and journalists runs on your customers every time they see your listing. The strikethrough price, the star score, the social proof badge — they're all setting anchors, and those anchors are shaping decisions before the customer reads a word of copy.
You can't turn the machinery off. You can only be deliberate about which numbers you feed into it.
The first number always wins. Make sure it's yours.