Pricing

Back when the thing being sold was a physical object, setting a price was close to arithmetic. If materials and labor for one screw came to 800 won, it sold for 1,200 won and the 400 won was visible. Selling more meant making more, making more cost more, and the cost of goods set a natural floor under the price.

Software sits in the spot where that floor disappeared. One more person using a document collaboration tool doesn't noticeably raise what the company spends. Server and storage costs go up by a few hundred won, so pricing from cost alone produces a product that costs 500 won a month. There is no basis for the number inside the product, so the basis has to come from outside it — what the customer feels the product is worth, and what other products solving the same problem charge.

When a team doesn't know the basis lives outside, the meeting ends in the same place every time. Pagedeck's first price plan was settled in a single meeting. A competitor charged 9,900 won a month, so the suggestion was to charge 9,900 won, nobody objected, and no document was left behind explaining why that number. Two years later, when Sofia asked where the price had come from, nobody in the room could answer.

Without a basis, every decision downstream wobbles. When an enterprise customer asks for 20 percent off, there is no standard for how far is too far, so each account manager answers differently, and the discounts written into contracts vary enough that eventually nobody can even compute the average amount actually collected. Adding thirty features and still not daring to raise the price has the same root. If nobody knows what the original 9,900 won rested on, nobody can calculate what the right number is now. The team ends up defending a figure it never chose, which is why the conversation keeps circling back to whether the founder would be comfortable with the change.

Pricing is the work of putting a basis under that number. The price is the point where money actually enters the business model, and it is also the sentence that declares how much the value proposition is worth. A business model is the structure of who you serve, what you give them, and how you make money; a value proposition compresses into one line the reason a customer should choose this product over the others. A product asking 3,000 won a month and a product asking 300,000 won a month are making different promises even with identical features, which makes the price a number and a message at the same time. The first promises something a person adds to a personal workflow and drops without much thought; the second promises something a company builds a process around and expects support for.

The five terms grouped under pricing split three ways. What the number leans on is the basis, how you check that the number matches the one in the customer's head is the research, and how the number is presented is the tactic. Walked in order, it becomes a flow: build the basis, ask to confirm it, then put it on the screen.

Basis is the branch that divides by where the price starts from, and three terms sit here. Each of the three looks somewhere different.

1. Value-based pricing — convert what the customer gains from the product into money and charge a slice of it. A team that spent four hours a week tidying documents and now spends one has gained three hours of labor cost, and if that comes to about 120,000 won a month, you ask for 30,000 of it. This earns the highest number of the three, but it takes the most work, because you have to find out what the customer saved before you can price against it.
2. Cost-plus pricing — add up what it costs to serve one unit of the product and put a margin on top. If server and storage for one account run 2,000 won a month, adding 30 percent gets you 2,600 won. The calculation is clear and easy to defend, but in software one more user barely adds cost, so using only this method sets the price absurdly low.
3. Competitor-based pricing — line up what other products solving the same problem charge and take a position next to them. If the competitor is 12,000 won a month, you either match, strip features and sit underneath at 7,000, or add and stand above at 20,000. It's fast when the market already has a reference point, but you inherit a number somebody else chose. If the competitor picked theirs without a basis, you inherit that missing basis too.

Research is the branch that checks whether the chosen number falls inside the range the customer has in mind, and one term sits here.

4. Van Westendorp — ask customers about price, but never "what would you pay?" Ask four things instead: at what price does it start to feel expensive, at what price is it too expensive to buy, at what price does it start to feel cheap, and at what price is it so cheap the quality looks doubtful. Overlay the four answers and one band remains with a ceiling and a floor, and the price gets chosen inside that band. The survey works because people can't name a fair price directly but answer the boundaries of expensive and cheap fairly consistently.

Tactic covers how the chosen number is shown, and there is one term here as well.

5. Price anchoring — set up a comparison before revealing the price. A 30,000 won plan standing alone looks expensive; put a 90,000 won tier beside it and the same 30,000 reads as reasonable. Nothing about the number changes, only how it is read. If the first four decide how much to charge, this one decides how that amount will look.

The five are not a list you pick one item from. In practice they stack: cost draws the floor, competitors confirm the rough band, what the customer gains measures the ceiling, a survey picks a point inside it, and the screen presents a comparison alongside. Cost is the lower bound below which selling loses money, value is the upper bound above which nobody buys, and the space between them is what there is to choose from.

Where inside that space to land gets narrowed by unit economics, the calculation that weighs what one customer pays against what it cost to acquire them. The lower the price, the thinner that gap. If acquiring a customer costs 50,000 won and the price is 3,000 won a month, it takes seventeen months to break even, and in a product where the average customer leaves after six, every sale deepens the hole. When that calculation doesn't work, the number is wrong no matter how carefully the other five steps were walked. The usual fix is not a bigger marketing budget but a higher price, a cheaper way to acquire, or a product people stay with longer.

"Is adding this one feature enough of a basis to raise the price?"

A proposal to measure how much the new feature actually raised what the customer gains, and also a way of saying that if it hasn't been measured, the price shouldn't move.

"I can go up to 10 percent on my own authority."

The discount ceiling is set in advance and inside it no approval is needed. The existence of a ceiling is itself a sign that the price has a basis under it.

"The problem isn't that we're cheap, it's that we look cheap."

A push to work on how the amount reads rather than the amount itself, which usually leads to changing the order the plans appear in or what they get compared against.

Pagedeck's plan overhaul started with one question from Grace: why had the price stayed at 9,900 won while the feature set doubled over two years? The first thing Mia did was not to name a new number but to go looking for the basis of the current one, and there was no basis to find.

So they split the measuring three ways. Mia asked twelve customer companies how much time they had cut from tidying documents and got an average of three hours a week. Ravi built a table of six competing products and found their plans clustered between 8,000 and 25,000 won a month. Sofia pulled the amounts actually collected from the last year of contracts and found that a nominal 9,900 won scattered between 6,200 and 9,900 once discounts were applied.

Laying the three numbers side by side changed the shape of the conversation. What had been a contest of instincts about whether to raise the price became a choice about where to stand, with a candidate of 14,000 won sitting between the hours customers save, the band competitors occupy, and the amount actually being collected today.

The most common belief is that a lower price sells more. Cheaper means more buyers is close to common sense, and looking only at signup counts it sometimes reads that way. But in software the price is also read as a signal of quality. To someone choosing where to put their company's documents, a product at 2,000 won a month doesn't look cheap, it looks untrustworthy. Halve the price, double the signups, and revenue holds flat while support requests and server load double — and a large share of those arrivals never intended to pay at all, so they drain away within months.

Treating it as a one-time decision goes wrong just as often. A pricing page gets built once and rarely touched, so the price looks like part of the furniture. But the product keeps changing and the market price moves with it. If features doubled over two years while the price held, customers are getting twice as much for the same money, and the difference is money the company let go past. That loss never appears on an invoice, so it accumulates without anyone noticing. Most teams revisit the price at least once a year and, when they decide not to raise it, write down why. Leaving the reasoning behind means the next person doesn't have to measure everything from the start again, and it turns a yearly argument into a yearly review of numbers that already exist.