# Scenarios and sensitivity

> Which assumption matters most, and a plan for the weak case.

- Course: [Finance for Product People: From Unit Economics to Business Cases](https://productovi.com/courses/finance-for-product.md)
- Module: What will next year look like?
- Lesson 26/32, ~28 min
- Source: https://productovi.com/courses/finance-for-product/scenarios-and-sensitivity

## What if we're wrong?

**Alex Tutor:** You're still on Klasna's product team, and the board has seen the forecast.

**Leo:** The board liked our 2.25 million. Then one of them asked: "And if you're wrong?" I said we'd be roughly right.

**Hanna:** Roughly right about which driver? They can't all hurt us equally. Some we could miss by a lot and barely notice.

**Alex Tutor:** That's **sensitivity analysis**: move one assumption at a time, keep the rest as they are, and see how far the result moves. Use the same step for each, say 10%, so you can compare them.

**Sofia:** Let me try price. 10% of 650 is 65 more per lesson, 63.375 after the 2.5% fee. Times 80,128 lessons: 5,078,112 UAH more profit.

**Alex Tutor:** More than twice the whole base profit, from one driver. Now the one on the other side of the lesson: what the tutor gets.

> **Exercise:** Keep everything else at the base case (80,128 lessons). By how much does the year's operating profit change if the tutor payout is **10% higher** than 415 UAH? Enter a fall as a negative number.

## Every driver, side by side

**Alex Tutor:** Here's every driver moved by 10% in the direction that matters, sorted by how much the year's profit moves.

> **Sensitivity of the year's operating profit (base: 2,251,968 UAH)**
>
> | Driver moved by 10% | Change in profit, UAH |
> |---|---|
> | Price per lesson +10% | +5,078,112 |
> | Tutor payout +10% | −3,325,312 |
> | Lessons per learner +10% | +1,500,434.25 |
> | Active learners +10% | +1,424,434.25 |
> | Marketing −10% | +309,000 |
> | USD rate +10% | −252,403.20 |
>
> The other drivers stay at the base case each time.

**Hanna:** Price and payout dwarf everything else. Our margin is thin: of 650, only 187.25 stays with us. Ten percent of a big flow beats ten percent of a small leftover.

**Sofia:** Why do more learners give less than more lessons per learner? Both add about the same number of lessons.

**Alex Tutor:** Because learners also drive coordinators. With 10% more, February and May cross a threshold and each needs one more: 76,000. More lessons from the same learners don't.

**Leo:** And I spent the whole planning week arguing about learner numbers with marketing.

> **Exercise:** Klasna's sensitivity table shows how much the year's operating profit (base: 2,251,968 UAH) moves when each driver alone moves by 10%. Where should planning time go first?

## How far can the price fall?

**Alex Tutor:** Now flip the question for the riskiest driver. How low can the average price go before the year's profit is zero, with everything else as planned?

**Alex Tutor:** Spread the fixed costs over the lessons, add what each lesson costs, and divide by what's left of the price after the fee.

$$
P^{*} = \frac{F / L + T + V \times FX}{1 - 0.025}
$$

*P*: break-even average price; F: the year's fixed costs (12,752,000 UAH); L: lessons (80,128); T: payout (415); V × FX: video + AI (31.50).*

**Sofia:** Fixed costs per lesson come to about 159. So the price has to cover roughly 606 of costs before the fee.

> **Exercise:** Use the formula: fixed costs 12,752,000 UAH, 80,128 lessons, payout 415, video + AI 31.50, fee 2.5%. At what **average price per lesson** does the year's operating profit reach zero? Round to two decimals.

**Hanna:** 621.17. That's only 4.43% below the plan and about where last year's average was. If the September price rise doesn't stick, the year is at zero.

**Alex Tutor:** Hanna's worry in numbers: what if the rise doesn't stick at all?

> **Exercise:** Klasna's base case: 80,128 lessons, average price 650 UAH, 2.5% payment fee, 2,251,968 UAH of operating profit. Families push back on the rise and the average price stays at **620 UAH**, with every other driver as planned. What is the year's operating profit? Enter a loss as a negative number.

## Scenarios: drivers move together

**Alex Tutor:** In real life drivers rarely move alone. **Scenario analysis** builds consistent stories instead. In a weak year fewer families come, more of them wait for discounts and the hryvnia slides: learners −10%, average price 630, the dollar at 45.

**Alex Tutor:** In a strong year: learners +10% and the dollar at 41, price as planned. Run each through the same sheet, month by month.

**Alex Tutor:** Before you run them: the board asked which assumption hurts most. Which of the two tools answers that?

> **Exercise:** A board member asks: "Which single assumption hurts us most if it's wrong?" Which analysis answers that question?

**Sofia:** Weak: −740,802 UAH. Strong: 3,742,508.

**Leo:** So the honest answer to the board is anywhere between losing 0.7 million and making 3.7 million?

**Hanna:** Yes. And the base case isn't the middle of that range. It's what we believe is most likely.

> **Exercise:** **True or false:** in Klasna's weak scenario (10% fewer learners than the base case, average price 630 instead of 650 UAH, the dollar at 45 instead of 42 UAH; the base case has 80,128 lessons and 2,251,968 UAH of operating profit) Klasna still delivers more than 72,000 lessons, yet the year ends with an operating loss.

## A plan for the weak year

**Leo:** I can't just tell the board we might lose money.

**Alex Tutor:** Then don't stop at the number. Agree now on a signal that tells you you're in the weak story, and on what you'll do when you see it.

**Hanna:** My proposal: if October's learners come in 10% under plan, we halve marketing for June to August, which saves 295,000, and postpone one of the two hires: 45,000 a month, 540,000 over the year.

**Alex Tutor:** Put Hanna's plan through the weak year before anyone says it out loud.

> **Exercise:** Klasna's weak scenario ends the year at −740,802 UAH. The weak-case plan: if October's learners come in 10% under plan, halve marketing for June to August (saves 295,000 UAH) and postpone one hire (45,000 UAH a month for 12 months). Nothing else changes. Where does the weak year end with the plan?

**Sofia:** Then the weak year ends at +94,198 instead of −740,802. Not great, but not a loss.

**Alex Tutor:** That's the real use of scenarios: a trigger and a response agreed in calm weather, not invented in a panic in July.

> **Exercise:** Which of these make a weak-case plan useful? Select all that apply.

## Your turn: the answer for the board

**Leo:** Write me the reply to "And if you're wrong?" Short enough to read out on a call.

**Hanna:** Which assumption matters most, the range, and what we'll do if the weak case starts.

**Sofia:** Driver, range, trigger and response. Got it.

> **Exercise:** Write Leo's answer to the board's question "And if you're wrong?" in 3–5 sentences: the assumption that matters most, the range of outcomes, and the plan if the weak case starts. Use the sensitivity table and these facts: base profit 2,251,968 UAH; break-even average price 621.17 UAH against 650 planned; weak scenario −740,802 UAH, strong 3,742,508 UAH; the plan (if October's learners are 10% under plan: halve summer marketing, postpone one hire) lifts the weak year to +94,198 UAH.

> **Key ideas**
>
> - **Sensitivity analysis** moves one driver at a time by the same step and ranks the drivers by how much the result moves.
> - With a thin margin, price and cost per unit dominate: +10% on price is +5,078,112 UAH, +10% on payout is −3,325,312, +10% on learners only +1,424,434.25 (step costs included).
> - A break-even value turns a driver into a red line: the year breaks even at an average price of 621.17 UAH, 4.43% under plan.
> - **Scenarios** move several drivers together in a consistent story: weak −740,802, strong 3,742,508.
> - The useful output is a plan: a trigger tied to a driver and actions agreed in advance (here +94,198 in the weak year).

**Alex Tutor:** Next time the teams turn this forecast into budgets, and we'll see why the numbers they send back are rarely the honest ones.

- Previous lesson: [Driver-based forecasting](https://productovi.com/courses/finance-for-product/driver-based-forecast.md)
- Next lesson: [Budgets people believe](https://productovi.com/courses/finance-for-product/budgets-people-believe.md)

---
Productovi Learning. Free, no sign-up needed to read.