Case: four ways to grow in 2027

A growth direction says what changes: the offer, the customers, both, or neither. Two questions sort any idea: do we sell something new, and do we sell to someone new?

• More of the current offer to current customers: same offer, same people, they buy more or earlier.
• A new offer to current customers: new offer, people who already know you.
• The current offer to new customers: same offer, people who've never bought it.
• A new offer in a new market: both change at once.

Klasna's four ideas for 2027 fall neatly into them. A (families already with Klasna start in October instead of January) changes neither offer nor customers, only how much they buy. B (Ukrainian language for current exam families) keeps the customers and changes the offer. C (adult English for Ukrainians in the EU) keeps the offer and changes the customers. D (Polish for adults moving to Poland) changes both.

Risk rises the further you move. D needs new tutors, new channels and customers who've never heard of Klasna, all at once, and each can fail on its own. Sounding bold isn't a reason: D faces the same tests.

The biggest number in a table is rarely the answer. Compare growth options on net, not on contribution alone:

likely net = likely contribution − cost.

Uliana's table gives each 2027 option a cost (marketing, hiring and training tutors, materials) and a low and a likely contribution (after tutors, video and payment fees). They're the team's assumptions, not facts.

• A: 300,000 − 120,000 = 180,000 UAH
• B: 600,000 − 450,000 = 150,000
• C: 480,000 − 360,000 = 120,000
• D: 500,000 − 700,000 = −200,000

B has the biggest likely contribution, but once its cost comes off, the smallest idea, A, has the best likely net in 2027. D loses money even in the likely case in its first year.

Two slips: comparing contribution without taking cost off, and reading one year as the whole story. A new subject or market may pay back more in later years; if you think so, say it and show the numbers for those years, rather than letting a first-year figure carry an unstated hope.

A likely number shows what you hope for. The worst-case net shows what you could lose:

worst-case net = low contribution − cost.

The low case in Uliana's table is a bad but plausible year, not a disaster movie. For the four 2027 options:

• A: 150,000 − 120,000 = +30,000
• B: 100,000 − 450,000 = −350,000
• C: 60,000 − 360,000 = −300,000 UAH
• D: 0 − 700,000 = −700,000

Only A can't lose money in a bad year. C, with a respectable likely net of 120,000, could cost Klasna 300,000; D could cost 700,000.

Show likely and worst case side by side for every option. The biggest likely number can hide the biggest possible loss: B's likely net is second best, but its downside is the second worst.

The slip is leaving the worst case out because it "probably won't happen". The point of pricing it is to ask whether Klasna could live through it, and to decide in advance what it would do if the year started going that way.

Once every option has a worst-case net, rank them by downside as well as by likely result. The two rankings rarely match, and the gap between them is where the real decision sits.

Klasna's 2027 options, from the smallest to the largest possible loss:

1. A, more per family: +30,000 (no loss even in a bad year)
2. C, adults abroad: −300,000
3. B, a new subject for current families: −350,000
4. D, a new market: −700,000

By likely net the order is A, B, C, D. B and C swap places: B promises more but can also lose more.

Now hold each downside against what Klasna can absorb. A small school that spent 2,850,000 UAH on marketing last school year can survive a bad year on A without noticing, and would feel C or B. D's worst case would eat a quarter of a year's marketing money.

The slip: treating the ranking as the decision. It tells you which bets are expensive to get wrong, so that a bigger bet comes with stronger evidence first, not that the cheapest bet always wins.

A is the only 2027 option that can't lose money in a bad year. It's tempting to conclude that the safest option is always the right one. It isn't.

Pricing the downside makes the choice honest: everyone sees what each bet could cost before agreeing to it. It doesn't make the choice for you. With cash to spare and strong early signals for B (say, many current families asking for Ukrainian-language help), a bigger bet could be right, because its likely result is large and its risk can be lived with.

What makes a riskier bet reasonable:

• the business can survive the worst case without hurting what already works;
• there's evidence, not only enthusiasm, that the likely case is likely;
• a leading signal will show early if it's going wrong, and a stop rule says what happens then.

The opposite slip is just as real: always picking the safe option and never learning anything new. A small team can make one main bet and still test the next idea cheaply on the side.

Every growth idea rests on a few conditions that must be true for it to work. Write them down before you spend, then check the riskiest one first.

For B, Ukrainian language for current exam families:

• Klasna can hire tutors who know the Ukrainian-language exam well;
• current families would rather take a second subject with Klasna than with a school teacher;
• the weekly progress note, Klasna's extra, works for the new subject too.

Two things are not conditions. Whether a large online school sells the subject isn't the test; families choosing Klasna is. And "Ukrainian lessons won't take any lessons from maths or English" is not required: some families will move hours from another subject. That's cannibalization, measured on total contribution, not a condition that must be zero.

Read A through the same lenses. Exam families who started by December 2025 paid 53,960 UAH on average; those who started in January 2026 or later paid 36,472. Starting earlier is more value for the family too. But A grows the exam segment, the most seasonal and contested one, while C and D build something year-round.

The growth directions are about new moves. An offer already in the range gets the stance its stage calls for, not a bet of its own.

Klasna Club is three months old and adults keep joining, but almost one in five joiners has already left, and in module 4 most July joiners who skipped the live clubs were gone within three months. So in the 2027 plan the Club:

• gets no growth bet of its own;
• keeps its job as the add-on for adults at today's price, 450 UAH a month;
• has its first month fixed first (module 4), before any push;
• has its stage read again once it has a year of history to compare with.

Two tempting mistakes. Making the Club the main 2027 bet with a big push and a lower price pours more people into a leaky first month and discounts adults who'd join anyway. Closing it because people cancel throws away a new offer that's still taking off.

Keeping the Club in view also matters for the growth choice itself: C (adults abroad) would bring adults who could join the Club too, which is a reason to check C cheaply.

A small team can't run four growth bets at once. One main bet a year is enough, plus a capped test that buys information about the next idea cheaply.

A capped test has three parts, written before it starts:

• a budget cap it can't exceed;
• a date by which it's judged;
• a stop or scale rule: what result keeps it going, and what ends it.

For example, a test of B: offer Ukrainian-language lessons to 30 current exam families in February and March, with a cap of 40,000 UAH, and continue in autumn only if at least five buy a package by 31 March. Without the cap, the date and the rule, "let's try it for a while" quietly turns into a second main bet.

For Klasna, that means A as the 2027 main bet and a capped test on the side.

How it fits together: the growth direction says how far a move goes, likely and worst-case net price it, the must-be-trues and a capped test check it, and the offer's stage and role keep the range in view. The next module asks whether marketing actually worked.