Module 4
Price it
A value metric, a price on the page, and unit economics that work at ten customers.
Pricing is the fastest lever you have and the one founders defer longest. Most early products are underpriced by a factor of three, which shows up later as a business that needs a thousand customers to survive.
This module produces a number you can put on the page, and the arithmetic behind it.
Sign in below to save your answers
Find the value metric
The value metric is the unit that grows as the customer gets more value. Vehicles inspected, invoices processed, tickets resolved, properties managed. It should be something the customer already counts, and something that rises when they are succeeding.
A good value metric makes the price feel fair as usage grows and makes expansion revenue automatic. A bad one is disconnected from value, which is how you end up with customers extracting a great deal more than they pay for, or leaving because a quiet month still cost them full price.
Write three candidate metrics and pick the one your customer would nod at without an explanation.
Your turn
What good looks like
- Candidates: per vehicle in the fleet, per inspection completed, per user account.
- Chosen: per vehicle. Operators already budget per vehicle per month, it grows as they grow, and it does not punish them for inspecting carefully.
Why per-seat pricing is breaking down
Per-seat pricing assumed software made a person more productive, so more people meant more value. When the software does the work instead of the person, that logic inverts: the better your product performs, the fewer seats the customer needs, and your revenue falls as your value rises.
The alternatives are per unit of work done, per outcome achieved, or a platform fee plus usage. Each has a failure mode. Pure outcome pricing invites arguments about attribution. Pure usage pricing makes revenue unpredictable and scares finance teams. A base plus usage is usually the practical answer for an early company.
Write which model you are using and the honest objection a buyer will raise.
Your turn
What good looks like
- Model: platform fee of $200 a month plus $2 per vehicle.
- Objection: 'our fleet size swings seasonally'. Answer: billed on average active vehicles, so a quiet month costs less.
Anchor to the cost you remove
Price against the customer's cost, never against your own. Your hours, your compute bill and your effort are irrelevant to the buyer, and anchoring on them is how founders arrive at prices that are a tenth of the value delivered.
Take the arithmetic from module one. If you remove $3,100 a month of cost, a price of $600 to $900 is easy to say yes to and still leaves you a real business. A price of $99 signals that the problem was not serious.
Write the value, the price, and the ratio. Anywhere between three and five times value to price is comfortable for a new company with no track record.
Your turn
What good looks like
- Cost removed: $3,100 a month for a 400-car fleet.
- Price: $200 base plus $2 per vehicle = $1,000.
- In the call: 'It costs you about a thousand a month and it takes about three off your dispute line. If it does not, cancel at the end of the month.'
Model unit economics at ten customers
Model ten customers, not a thousand. A model at a thousand customers is fiction and hides everything. A model at ten is checkable and tells you whether the business works.
Five numbers: price per customer per month, direct cost to serve one customer including compute and support, gross margin, what it costs you to acquire one, and how many months until you get that back.
If the answer at ten customers is that you lose money on every one and make it up on nobody, the problem is the price or the cost to serve, and it does not fix itself with scale. Fix it here, where changing your mind is free.
Your turn
| Metric | Value | How you got it |
|---|---|---|
What good looks like
- Price: $1,000/month. Cost to serve: $180 (storage, inference, support hours). Gross margin: 82%.
- Cost to acquire: about $900, mostly my time at a notional rate.
- Payback: just over one month. Ten customers = $8,200 a month of gross profit.
The judgment call
The number on the page is a judgment call. The arithmetic narrows the range, it does not choose the price. No further analysis will tell you whether to be at the top or the bottom of your range, because the answer depends on how much risk you want the first ten customers to absorb.
How to think about it
- Price for the customer you want in a year, not the one you are nervous about losing this week.
- Discount on term or volume, never on the list price. A quiet discount today becomes your permanent price.
- If nobody has flinched at your price, it is too low. About one in three should hesitate.
- Decide now what you will do when a good customer asks for half off.
Questions worth sitting with
- What does the buyer spend today on the thing you replace, including the parts they do not label as cost?
- Which customers would you be relieved to lose at a higher price?
- What happens to your business if you charge 50% more and close a third fewer deals?
- Are you pricing low because of the market, or because you are not yet comfortable asking?
Common questions
How should an early startup set its price?
Anchor to the cost you remove for the customer, choose a value metric they already count, and aim for a value to price ratio of roughly three to five. Never price from your own costs or hours.
Is per-seat pricing dead for AI products?
It is breaking down wherever the software replaces the work of a seat, because your revenue then falls as your product improves. A platform fee plus a usage metric is usually the practical alternative.
Why model unit economics at ten customers?
Because a ten-customer model is checkable and a thousand-customer model is fiction. If the margin and payback do not work at ten, scale will not repair them.