The value metric: what the price grows with

Every price is charged per something: per lesson, per month, per seat, per minute, per result. That unit is the value metric. It decides who pays more and who pays less as use grows, so it matters as much as the number on the price tag.

Klasna already uses two. One-to-one lessons are sold per lesson: each lesson gives the family a lesson's worth of value and costs Klasna a tutor's payout, so price, value and cost move together. Klasna Club is sold per month at 450 UAH, with unlimited use of the AI practice partner. A month is simple and easy to budget, but it treats a member who practises 20 minutes the same as one who practises 20 hours.

The value metric comes before tiers and discounts. If the unit is wrong, no price level fixes it: the wrong people end up paying too much or too little.

Before choosing a unit, check what one more unit of use costs you. The Club's AI practice partner is billed in dollars per minute: 0.80 UAH a minute at the planning rate of 44.8 UAH per USD. The speaking-club hosts cost 55 UAH a member a month, whatever the member does.

So one monthly member's contribution, what's left after the costs that come with that member, is:

450 × 0.975 − 55 − 0.80 × minutes

The 0.975 takes off the 2.5% payment fee and leaves 438.75; minus the hosts, 383.75 before any AI. The average member practises 50 minutes, costs 40 UAH of AI and leaves 343.75. That average is why the Club's direct cost is quoted as 95 (55 + 40).

But averages hide people. A member at 400 minutes leaves 383.75 − 320 = 63.75 UAH; one at 900 minutes leaves −336.25. With a flat price and a cost per use, every member has a different contribution, so work it per member, not per average.

With a flat price and a cost per use, every member has a point where they stop paying for themselves: the usage at which contribution reaches zero. Divide what's left before usage by the cost of one unit:

break-even minutes = (price after fee − fixed per-member costs) ÷ cost per minute

For a monthly Club member: 383.75 ÷ 0.80 = 479.69 minutes, about eight hours of practice a month. The annual plan works out at 360 a month and leaves less: (360 × 0.975 − 55) ÷ 0.80 = 296 ÷ 0.80 = 370 minutes. Cheaper plans reach the point sooner.

Nazar's reply, “most people do a fraction of that”, is true and beside the point. «Unlimited» attracts exactly the few who do more, and the dollar-billed cost grows with every one of them. If the hryvnia weakens, the cost per minute rises and the point drops further. Treat it as a risk marker: what matters is how many members sit above it.

“It's software, another minute costs nothing” is true for some products and false for others. An extra click in a simple web tool really costs close to nothing. An AI feature is different: every answer runs a model that someone bills per use, here in dollars per minute. Card payments (a fee on every charge) and human support also cost per use.

So check the marginal cost of a feature, the cost of one more use, before promising it «unlimited». If it's near zero, unlimited is a fine selling point. If it isn't, unlimited hands the heaviest users a bill that you pay: at 0.80 UAH a minute, a monthly Club member above 479.69 minutes costs Klasna more than the 450 they pay. The risk also moves with the exchange rate, because the bill is in dollars and the price is in hryvnia.

None of this means removing the feature. It means the price needs some link to use, which is the job of the value metric.

A good value metric passes three tests.

1. It grows with the value the buyer gets. Someone who gets more out of the product pays more, which feels fair to both sides.
2. It grows with your cost to serve. Otherwise heavy users drift into loss while light users pay for them.
3. The buyer can predict the bill. People budget; a unit they can't foresee makes them hesitate or ration use.

Klasna's per-lesson price passes all three: a lesson is the value, a lesson triggers the tutor's payout, and a family knows what a 16-pack costs. The Club's flat month passes the third test, the first only roughly, and fails the second for heavy users of the AI partner.

Two tempting shortcuts aren't tests: “easy for us to count” (logins are easy to count and say nothing about value) and “what rivals use” (their costs and buyers may differ).

Run the candidates through the three tests (grows with value, grows with cost, predictable bill).

• Per month, unlimited AI: simple and predictable, but the AI cost grows per minute, so the heaviest users can cost more than they pay.
• Per practice minute: matches cost perfectly but hurts learning. A learner watching a meter practises less, and practice is what the Club sells. Members also can't predict the bill.
• Per result (pay when the speaking level goes up a step): matches value on paper, but the result depends on the learner too, «a step up» needs a test both sides accept, and disputes eat the margin. Use results in what you promise and measure, not as the meter.
• Per lesson: few weaknesses where it fits. Value and the tutor's payout grow with each lesson.

No unit wins every test. The choice is which weakness you can manage, and for the Club that points to a mix: a monthly price with a limit on the costly part.

When a flat price suits buyers but one part of the product has a cost per use, the usual answer is a flat price with an included amount. The month includes a stated quantity of the costly part, and there's a clear, fair way to get more: a top-up or a higher tier.

Why it works: typical members never meet the limit, so the price stays simple and predictable and still roughly tracks value. The limit caps the cost of the heaviest users, and the top-up turns heavy use into revenue instead of a loss.

For the Club, that means a monthly price with a number of AI minutes included, set well above what most members use (the average is 50 minutes) and below the point where a member stops paying for themselves (about 480 minutes on the monthly plan, 370 on the annual). This module's case sizes the exact numbers from the Club's usage data.

A limit is fair only if buyers see it before they pay. Writing the included amount on the price page, next to the price, lets people judge the offer. Discovering it later, when the AI partner stops in the middle of a conversation, feels like a broken promise and invites complaints and refund requests.

Three rules keep an included amount honest:

• Generous: sized so typical users never notice it, not set to squeeze the average member.
• Visible: the number, what happens when it runs out and the price of more, all on the price page.
• Fair to current members: a change of terms is announced in advance, with time to decide, not switched on overnight.

How it fits together: the value metric is the unit the price grows with; cost per use and the break-even point show when a flat month stops working; the three tests pick the unit; a stated included amount closes the gap. Next: tiers, a cheaper Club, Club Plus and Dana's three offers.