Pricing is a positioning decision.

The pricing meeting has been running for ninety minutes and it is still about the number. Someone wants 15% more because the product has improved. Someone wants to hold, because the last two losses cited price. A third person has built a model showing both are defensible, which is true and unhelpful. The room is arguing about where to put a figure on a page.

That is the wrong object. A price is a position in a set the buyer assembled before they ever saw your page — and the set is the thing you can actually move.

The uncomfortable part is that the buyer’s sense of what your category costs was substantially settled before the conversation started. The useful part is that it was settled by comparison, and comparison is something you can supply.

None of which is a new claim. April Dunford has spent a decade teaching B2B teams to start from the alternatives a buyer would otherwise pick, and “your pricing problem is a positioning problem” has been somebody’s headline more than once. What has been missing is a measurement. If the reference set really is assembled upstream, that should leave a trace in what people search for — and search volume can be counted. So I counted it, in three categories across two markets, and the ratios came back further apart than the argument needs them to be.

Valuations are built on the spot

This is one of the few commercial claims with experimental evidence sitting directly underneath it — studies built to test this exact question.

Dan Ariely, George Loewenstein and Drazen Prelec ran a series of studies on how people arrive at what something is worth to them, and published the result in the Quarterly Journal of Economics under a title that gives away the finding: “Coherent Arbitrariness”: Stable Demand Curves Without Stable Preferences. Their opening claim:

In six experiments we show that initial valuations of familiar products and simple hedonic experiences are strongly influenced by arbitrary anchors

Two words in that sentence do the work. Familiar — these are not exotic goods where you would expect people to be lost. And arbitrary — the anchors had no informational content whatsoever.

The second half of the paper’s title is the half that matters commercially. Preferences are not stable, but demand curves are: once an initial valuation is set, later valuations stay coherent relative to it. People will reliably pay more for more, and less for less, around a starting point that was itself accidental.

For anyone setting a price, that is a precise instruction. Arguing a buyer up to an absolute level is working against the grain. Deciding what the buyer’s starting point is, and what sits next to you once they have one, is working with it.

What sits beside you changes what you are worth

The mechanism was described a decade before the coherent-arbitrariness work that made anchoring famous. Itamar Simonson and Amos Tversky set out two effects in the Journal of Marketing Research, and the abstract states them plainly:

The first hypothesis, tradeoff contrast, states that the tendency to prefer an alternative is enhanced or hindered depending on whether the tradeoffs within the set under consideration are favorable or unfavorable to that option. The second hypothesis, extremeness aversion, states that the attractiveness of an option is enhanced if it is an intermediate option in the choice set and is diminished if it is an extreme option.

Read the second one as a pricing instruction and it stops being psychology. An option’s attractiveness changes depending on whether it is in the middle of the set or at the end of it — and whether your offer is in the middle or at the end is decided by what you put beside it.

This is the mechanism underneath a piece of pricing advice so well-worn it is nearly furniture — offer three plans. Everyone has heard the rule. The reason is the part that tells you which three. A single price is an extreme by construction: it is at once the floor and the ceiling of the page, and the buyer has nothing to do with it except accept or reject. Three plans convert an accept-or-reject into a which-one, and they move your intended offer from the end of a set to the middle of one.

The authors are careful about scope and so should we be. This is experimental work with consumer products, published in 1992, and a procurement committee running a formal evaluation is not a person choosing a camera. The direction transfers; the magnitude does not, and anyone quoting a percentage uplift from this literature into a B2B deal is inventing it.

You can watch the set being built

Both papers are laboratory work, and it is fair to ask whether buyers behave this way outside one. There is a way to look. Search demand is the visible residue of someone assembling a comparison set, and it can be counted.

So I counted it: twelve months of Google search volume for three B2B software categories in the United States and Germany — the bare category term, the comparison-shaped variants of it, and the price-shaped ones.

Comparison-shaped searches — “best”, “top”, “comparison”, “reviews”, and the German “Vergleich” and “Test” — are not much bigger. They run at roughly 3% of category volume in the three US categories, and wider and noisier in the German ones, where the absolute numbers are small enough that Google’s rounding does real damage. The German and English modifier sets are also not exact translations of each other, so I would not lean on the cross-market difference; the within-market shape is what holds.

And that shape says something specific. The moment the set gets assembled is the category query itself. Someone types two words, meets a page of listicles, review platforms and whoever bought the ad, and leaves with a reference set already formed. Your price page is not in that moment. It gets visited later, by a person who arrives already knowing roughly what this kind of thing costs — and the anchor they measure it against was set on a page you do not own.

Which is the practical version of coherent arbitrariness. The absolute level came from somewhere upstream and largely accidental. What you control is where you sit relative to it.

What this changes about the work

The practical consequence is that the decisions determining your price happen outside the pricing meeting.

The first row is the expensive one, and it is the one that has been tested directly. Brenda Kuijken, Gerda Gemser and Nachoem Wijnberg examined what category membership does to willingness to pay and reported in the Journal of Product Innovation Management that “consumers’ willingness to pay for a product under consideration will tend toward the average price of products in the category in which it is positioned” — the category average operating as the reference point. Their setting is consumers and new products rather than a procurement committee, so read it for direction only. But the direction is the whole row: filed differently, valued differently, with nothing about the product changed. A buyer in a new market reaches for the category average in front of them rather than the one you spent years building at home, which makes the level the part of a price that does not cross a border.

If the buyer files you under a category with an established price band, you are working inside someone else’s anchor, and the discount conversation was lost before it started. Changing the number is a fight; changing the filing is a positioning decision, and it is the one that moves the band.

That is also the connection back to why some commercial work accumulates and some resets. A price anchor is a stock. It was built by everything the market has seen from your category and from you, it does not respond to a quarter of effort, and publishing a new page does not reset it. The price page only reports it.

The objection

Someone reasonable will say that this is all fine for a consumer choosing between cameras and beside the point for an enterprise buyer running a structured evaluation with a scoring matrix and three references.

Partly right, and the part that is right is worth conceding properly: a formal procurement process is explicitly designed to suppress exactly these effects. Weighted criteria, blind scoring and a written business case all exist to stop the decision being made by whatever happened to be adjacent.

But the reference set that procurement scores against was assembled by someone, earlier and informally, and the scoring criteria were written by a person who already had a mental model of what this kind of thing costs. The evaluation constrains where the price lands within a set it inherited. And the shortlist — three vendors, one of whom is there to make the other two look reasonable — is a choice architecture whether anyone designed it deliberately or not.

What to do about it this week

Ask your last five won deals and your last five lost deals one question: what else was on the list?

Not “who did we compete against” in the CRM field sense, which records the vendor who made it to the end. The actual question is what the buyer had in their head when they started — including the options that never became a deal: doing nothing, building it internally, extending a tool they already own, hiring someone.

That list is Dunford’s. She calls them competitive alternatives, and the discipline of writing down what a buyer would do if you did not exist is hers. What is added here is the inference she does not draw: those alternatives are also the reference points your number gets read against, which means the same exercise that fixes your positioning also tells you your price band.

Write the answers in one column. Then, beside each, write what that option costs.

That column is your price band: the set of things your buyer was already weighing when you arrived. If the answers are inconsistent across ten deals, you do not have a pricing problem. You have a positioning problem wearing one, and work on the number will not fix it, because there is no single set for the number to sit in.

If the answers are consistent and you are at the top of the band, you are an extreme option. If you are in the middle, someone put you there, and it is worth knowing whether it was you.

The reason the pricing meeting keeps running long is that it is being asked to decide something it does not control. The number on the page is the last and smallest decision in a sequence that started with what the buyer thinks you are — and by the time anyone opens the spreadsheet, the answer has largely been given by people who were working on something else.