How to Set a Price Before You Have a Product

There is a distinction that clears up most pricing arguments in ten seconds. A pricing metric is a unit of measure: consume more, pay more. A value metric is the unit of consumption that actually creates value for the customer (Oh Blimey). You want the second one, priced.

Worked examples of the same product with different units:

• HubSpot prices some plans by number of marketing contacts (Oh Blimey).
• Airbnb charges for nights booked (Oh Blimey).
• A care software product can pick between active contacts, workflows completed, new organizations added or alerts recorded, depending on which one tracks the customer's progress (Oh Blimey).

That last list is the useful one, because four plausible units for one product is normal. The right unit is the one where growth in the customer's own success shows up as growth in the number.

A test you can run in five minutes. Ask: if the customer doubles this number, is it because they are getting more value or because they are getting sloppier? Seats pass when collaboration is the value. Seats fail when one power user does all the work and invites spectators. Events pass when each event is a unit of work the customer would otherwise do by hand. Storage fails almost always, because storage grows when people are disorganized.

The US Chamber's pricing overview is a useful map of the standard models, including flat rate, pay-as-you-go, tiered, per user and subscription (US Chamber of Commerce). Treat that list as a menu of shapes, then attach your chosen unit to one of them. Most pre-product B2B products land on a per-unit subscription: a monthly base plus a band of units.

Write your chosen unit in one sentence on the plan: "Priced per active workflow completed per month." If you cannot finish that sentence, you are not ready to interview.

Ten is a working number, not a law. It is enough to surface a pattern and small enough to do before you build. The sessions have one job: find out what one unit is worth to a qualified buyer.

Qualify hard first. A qualified buyer has the problem, has a budget line that could absorb the purchase, and has authority or access to whoever does. Maurya's guidance on customer segments is to distinguish customers from users, define the early adopter rather than the mainstream buyer, and keep multiple perspectives in view (Ash Maurya). Apply that filter before you book anything.

Then run this script, in order:

1. Ask them to walk you through the last time they did this task by hand. Get the hours and the people involved.
2. Ask what that hour or that person costs them, or what they currently pay an alternative to avoid it.
3. Name your unit and ask: if one unit of this were handled for you, what would it be worth per month?
4. Ask what they would expect to pay, then stop talking and let the silence work.
5. Ask what would make them say no at that number.
6. Ask who else would have to approve the spend.

Step 3 is the one people skip. Asking what the unit is worth, before naming a price, keeps the anchor on their value rather than your guess. If they cannot answer, ask them to compare it to the closest alternative they already pay for: a contractor hour, a tool seat, an agency retainer.

Convert everything to one number. If a buyer says an hour of the task costs them a loaded rate of $60 and your unit removes two hours a month, one unit is worth about $120 a month to them. Write the arithmetic down per interview. Ten interviews give you ten numbers with their reasoning, which is more useful than ten opinions.

Value-based pricing is defined as setting rates by what buyers believe the product is worth rather than by production cost (US Chamber of Commerce), and a pricing primer for product managers makes the same point: price proportional to the differentiated value over the next best alternative, as perceived by the customer (Medium). The interviews are how you get to "perceived."

Two hygiene rules. First, do not run the interviews as a group exercise with your team in the room; Maurya's canvas guidance warns that group sessions invite groupthink, and the same applies to customer calls (Ash Maurya). Second, record the numbers in the buyer's words, not yours. "That is worth about two hours of my analyst's week" is evidence. "They seemed positive" is not.

Sort the ten value estimates. You are looking for the 30th percentile and the 80th percentile, not the average. The average hides the shape of the market.

A worked example with numbers, all invented for illustration:

• Ten estimates of monthly value per unit: $40, $55, $60, $70, $75, $80, $90, $110, $150, $300.
• 30th percentile: about $62. 80th percentile: about $150.
• Capture ratio of 10 to 20 percent of stated value, a common starting band for a new tool with no track record. Floor: about $6 to $12 per unit per month. Ceiling: about $15 to $30.
• Round the floor to a clean published number, say $9 per unit per month, and set the ceiling at $25.

The floor is what you publish. It is low enough that a qualified buyer can say yes without a procurement fight, and it is honest because your stated value support is roughly $60 and up. The ceiling is internal. It is the number you move toward as usage data and references accumulate, and it tells you how much room the model has.

Why a band instead of a number? Because a single number implies precision you do not have before the product exists, and because the first ten conversations will not survive contact with the first hundred. A band also protects you from the most common early mistake in the other direction: giving the product away for free to people who would have paid (Oh Blimey). If you have a floor written down, you have a defensible reason to charge it.

Check the floor against your cost side before you commit. Add up variable cost per unit and confirm the floor clears it with margin. The US Chamber's step-by-step guide starts there, with variable costs per product, before any markup or value conversation (US Chamber of Commerce). A floor below variable cost is not a strategy, it is a subsidy.

The reason prices drift badly is that nobody agreed in advance what would justify a change. Fix that on the same page. Write three lines: the metric, the threshold and the action.

A workable default, and the one this article recommends: change the price when more than 20 percent of qualified buyers choose the same option across ten consecutive conversations. If eight of ten buyers in a row pick your top option, either the top option is underpriced or the tiers are shaped wrong. Both are fixable, and both need a decision rule rather than a feeling.

Triggers worth writing down, each with its own threshold:

• Selection concentration: over 20 percent of qualified buyers pick the same option across ten consecutive conversations. Action: re-cut the tiers or raise that option's price toward the ceiling.
• Conversion with no price objection: ten qualified buyers convert with zero questions about price. Action: raise the floor by one band.
• Objection clustering: most losses cite the same feature gap, not the price. Action: hold the price and fix the product; do not discount.
• Usage above band: most accounts sit in the top band of the unit within two months. Action: add a band and move the ceiling.

Set the review cadence at the same time. Quarterly is enough. The point of the cadence is not to change the price every quarter, it is that the conversation happens on a schedule instead of during a tense sales call.

Two more rules keep the trigger honest. Price dictates perception: a price that is too low can read as low quality, and a price that is too high is hard to sell without matching substance (Convert). And packaging is part of the price: bundling supporting material with a service raises perceived value more than a bare item does (Convert). So when a trigger fires, consider re-packaging before you consider a raw increase.

State this on the page in plain language: the first cohort keeps its original price for twelve months from signup. Three reasons, all practical.

First, it makes the sales conversation easy. You are asking people to buy something that does not exist yet, and a written price guarantee lowers the risk on their side of the table. Second, it buys you permission to raise prices later without breaking trust, because the promise had a date on it. Third, it forces you to plan the migration, which is the part most teams forget until the renewal email.

A grandfathering note you can adapt:

Early access pricing
- Unit: per active workflow completed per month
- Your price: $9 per unit per month
- Locked for: 12 months from your signup date
- After 12 months: your account moves to the then-current plan
  unless you tell us otherwise at least 30 days before renewal
- Reason for any change: published on this page

The 30-day notice line matters. It turns a surprise into a decision, and it gives you a clean test of who actually values the product.

Grandfathering has a cost, so size it. If your first cohort is small, the revenue forgone over twelve months is small and the reference value is large. If the cohort is large, cap the offer by date rather than by count: "signups before the public launch date." A date is enforceable and easy to explain.

The one-page plan needs the pricing line to be readable in ten seconds. Keep the public version to four lines. Everything else, including the ceiling and the interview notes, stays in a working document.

The Lean Canvas is a good place to keep the internal half, because its revenue streams box sits next to key metrics, and the pairing is exactly what a pricing decision needs (Canva). Fill the canvas per the standard nine-box structure described by IdeaScale, and use the key metrics box to hold your unit and your trigger (IdeaScale). If you prefer the broader nine-block structure, the Business Model Canvas and its block-by-block questions are laid out by Conceptboard, including the question that matters here: what are our customers willing to pay for this value (Conceptboard).

One note on tools. Lean Canvas is a snapshot, and Maurya's own guidance is to sketch it in under 20 minutes rather than perfect it, because the goal is a snapshot and not a prediction (Ash Maurya). Apply that to pricing too. The first numbers are a snapshot with a decision rule attached. Miro's comparison of the two canvases is useful if you need to pick one for a mixed audience of operators and investors (Miro).

If the writing and building is the bottleneck rather than the thinking, you can describe the plan, the page and the first version of the product and have a specialist pick it up and build it with you at Quanta.