Case: Nazar's price recommendation for 2027/28

A price recommendation starts from a simple trap: the highest price always looks best per sale. Whether it is best depends on how many people buy at it, and how many come back.

Leo wants 720 for the English 16-pack: with next year's costs (payout 410, video and AI 31.47, the 2.5% fee), one lesson would leave 720 × 0.975 − 410 − 31.47 = 260.53 UAH, against 211.78 at 670. Per lesson, yes, if families bought it.

The test says they bought it less and came back less. In the weeks when 720 ran, it converted 5.25% against 7.00% at 620, and only a third of its buyers returned within 60 days. And it ran only in the exam peak, never in a normal month. A page has to rest on what the evidence supports, so margin per lesson is one input, weighed against conversion, retention and what was actually tested.

A recommendation that raises a price should say how much of the rise is costs and how much is value. That split shapes both the decision and the message.

Start from the cost-neutral price: the price that keeps today's contribution per lesson after the new costs. For the English 16-pack it is 642.02 UAH, against 620 today. So of a +50 step to 670, 22.02 UAH is costs: 22.02 / 50 = 44.04%. The other 27.98, a bit over half, has to be justified by value: the spring price test showed families buying almost as often at 670 as at 620, with the same retention.

The split keeps the announcement honest. "Because of inflation" would overstate the cost story; "tutors are paid more, part of our costs is in dollars, and we keep investing in what you get" says both reasons, in proportion.

"Let's just keep the price" feels like no decision at all. It isn't: when costs rise, holding the price is a choice to earn less on every sale.

At 620, with next year's costs (payout 410, video and AI 31.47, the 2.5% fee), an English 16-pack lesson leaves 620 × 0.975 − 410 − 31.47 = 163.03 UAH, against 184.50 today. Holding means about 21.47 less on every lesson sold, across thousands of lessons a month.

Sometimes that's the right call: when demand is fragile, when a rival is pulling buyers away, or when the evidence for a higher price is weak. But a recommendation should name it as a choice with a cost, not present it as the safe default. Klasna's evidence points the other way: the test showed families accepting 670 with no clear loss.

Put the number on the page, so the option is compared fairly with the others.

Every option on the page needs the same yardstick: contribution per lesson at next year's costs. Only then can holding, cost-only and a value rise be compared fairly.

With payout 410, video and AI 31.47 and the 2.5% fee:

• 620 (hold): 163.03 per lesson;
• about 642 (costs only): 184.50, today's level kept;
• 670 (+50): 670 × 0.975 − 410 − 31.47 = 211.78;
• 720 (+100): 260.53, but tested only in the peak.

The recommended 670 leaves 211.78 a lesson, more than today, and it is the price the test supported: more revenue and contribution per visitor than 620, with the same retention.

Showing all four on one yardstick does two things. It makes the trade-offs visible to Leo and Hanna, and it shows that the recommendation is a choice among options, not a number picked first and defended afterwards.

A recommendation makes a choice. Lay out the options, pick one, and say in a line why each of the others loses.

For Klasna's English 16-pack:

• Hold 620: it converted best in the test, but conversion isn't the goal; at the new costs a lesson leaves only 163.03.
• About 642, costs only: keeps today's contribution but leaves unused what the test showed families accept.
• 670 (+50): tested, more revenue and contribution per visitor than 620, the same retention, covers the costs. Recommended.
• 720 (+100): its high overall revenue per visitor came from running only in the exam peak; in the same weeks it was lowest and kept fewer buyers.

The pattern works for any price decision: the evidence for the chosen option, and the specific reason each alternative falls short.

A recommendation is more trustworthy when it says clearly which parts rest on evidence and which on judgment.

Klasna's test covered the English 16-pack only. Nazar proposes one +50 step on every package, maths and programming included. That is simple to explain to families, and probably right, but for maths and programming it's a judgment, not a test result. The page says so: "tested for English; for maths and programming the same step is our assumption, and we will watch their sales closely from June".

Naming an assumption isn't weakness. It tells the readers where the risk sits, what to watch, and what would make you change course. Hiding it does the opposite: if maths sales drop, nobody knows whether the plan failed or the assumption did.

The same applies to costs: the 47.0 UAH per dollar behind the dollar line is a planning assumption, not a forecast.

The page should fix, before the change, how its result will be read. Otherwise the first loud month decides for you.

Klasna's plan: notice on 1 June, packages bought before 1 September at the old price. That means families will buy ahead in summer, cash will arrive early and lessons come later. So the page says:

• The window and the yardstick: June to December against the sales trend, never September against last September alone.
• Pre-buying is expected and planned for: tutor hours in autumn, cash not spent as if lessons were delivered.
• What would make Klasna reverse: for example, sales in the window clearly below the trend, or enquiries lost to price rising sharply.

What the page doesn't promise is to match rivals' prices: Klasna decides on its own costs and its own families. Writing all this down in May is what protects the decision in September.

A one-page recommendation is easier to trust when it states its scope: what it decides, and what it leaves to other decisions.

Nazar's page covers Klasna's hryvnia packages for 2027/28. Two lines handle the rest. The rival at 540: "we watch lost enquiries and differentiate; we don't match, and we don't discuss prices with them". Out of scope: the EUR prices for families abroad weren't in the test and get their own review on the same date; Klasna Club has its own pricing decision.

A page that says what it doesn't cover is easier to trust than one that seems to cover everything, and it stops readers from applying its evidence where it was never gathered.

How it fits together: that's the whole course: a floor and a ceiling, what buyers compare and pay, what you sell and how you show it, discounts, changing prices, tests and a yearly review. Whatever you price next, start from the evidence and say your reasons out loud.