The pricing review

Prices tend to change for the wrong reasons: a loud month, a rival's ad, a manager's hunch. A yearly pricing review replaces those reactions with one planned decision, made on a fixed date with all the evidence on one page.

Klasna's review happens every November and answers six questions:

1. What did we actually get per lesson (the realized price)?
2. What moved sales: price, volume or mix?
3. Have costs moved the floor?
4. What do competitors charge?
5. What did tests and customers tell us?
6. The decision: hold, change or test.

One page, not a 40-slide deck nobody reads. And "hold" is a full answer when the evidence says so: the point isn't to change something every year, but to make the decision on purpose, once a year, instead of in a panic after one bad month.

When sales change, the first question of the review is why. A sales change can be split into three parts: price, volume and mix. Start with the price effect: what the price change itself earned on the units actually sold.

Price effect = units sold this period × change in price per unit. If the price changed differently for different products, do it per product and add up.

Klasna's October English sales were 4,756,800 UAH in 2027 against 4,256,800 a year earlier: +500,000. Leo says it's all the price rise. The September change added +50 to every package size, and Klasna sold 7,184 English lessons this October. So the price effect is 7,184 × 50 = 359,200 UAH.

Use this period's units, not last year's: the price effect asks what the new price earned on what was actually sold at it. Whatever is left of the 500,000 comes from selling more lessons, or different ones.

The second part of the split is the volume effect: what the extra (or missing) units would have brought at last year's average price.

Volume effect = (units now − units then) × last year's average price per unit.

Last October Klasna sold 6,960 English lessons for 4,256,800 UAH: an average of 4,256,800 / 6,960 = 611.61 per lesson. This October it sold 7,184, so 224 more. Volume effect: 224 × 611.61 ≈ 137,000.46 UAH.

Using last year's price here keeps the two effects apart: the price effect already counted the +50 on this year's lessons, so the extra lessons are valued as if nothing about the price had changed. If you valued them at the new price, part of the price effect would be counted twice.

Volume can grow for many reasons: a trend that was already there, marketing, a good season. The split says how much, not why.

After price and volume, what remains of the sales change is the mix effect: buyers choosing a different combination of products, at different prices per unit.

Mix = total change − price effect − volume effect.

For Klasna's October: 500,000 − 359,200 − 137,000.46 = 3,799.54 UAH. Packages have different prices per lesson (the 8-pack is dearest per lesson, the 32-pack cheapest), so when families shift between sizes, the average price per lesson moves even with no price change. Here the shift was tiny and slightly toward dearer-per-lesson packages.

So the answer to Leo: about 359,000 is price, about 137,000 is volume and under 4,000 is mix. He's mostly right, but not all of it is the price.

A large mix effect is worth a closer look: it can mean a pricing change pushed buyers toward bigger or smaller packages, which affects cash and capacity too.

A price–volume–mix split explains sales. It says nothing directly about profit, because costs may have moved at the same time.

At Klasna they did: tutor payouts rose from 390 to 410 per English lesson in September, and the dollar costs follow the planning rate. So the 359,200 of price effect is not 359,200 of extra profit. Part of each +50 pays the higher costs; about 22 of it was the cost increase in the first place.

The review only needs the direction and the size of each effect, and an honest note on what the split doesn't show. The contribution view, and the full variance method that splits costs too, live in the course Finance for Product People.

Write it that way on the page: "sales +500,000, of which price about 359,000; contribution is affected by the new payouts, see the finance view". That keeps anyone from reading a sales split as a profit figure.

AI tools can take the routine out of a pricing review. Nazar's assistant reads twelve competitor price pages every week, summarizes support chats that mention price and drafts the review page. That's sound use: monitor, summarize, draft, with a person checking every result.

Three rules keep it safe:

• Personal data stays out. No families' names or messages go into an outside tool unless they are stripped of personal data first.
• Synthetic "customers" are not evidence. Asking a tool to play 200 parents and answer a price survey produces text, not what anyone will pay.
• The tool never sets a price for a person on its own. A personal price per family from their browsing is unfair to families and, in the EU, must be disclosed (as of October 2026).

People decide; the tool saves them time on the reading and the first draft.

An AI summary can be confident and wrong. It reads fast, but it can miss a condition, mix up units or take a number from the wrong line. So every number that goes into the review is checked against its source by a person.

This week Nazar's assistant reported that one rival cut English to 480 UAH. Opening the page showed the catch: 480 is per person in a group of four, not a one-to-one lesson. Put into the review as it was, that number would have suggested Klasna's 670 was far above the market and could have pushed a needless cut.

The check is quick: open the page, find the number, confirm what it's for (one-to-one or group, per lesson or per month, which package). Keep the link next to each competitor price on the review page, so anyone can verify it later.

The same review works for a one-person business, scaled down. Dana will do hers every August, before her new school year starts. Each question has its own evidence:

• Does the rate pay for my time? Pay per hour worked against an hour at Klasna; her rate against her target of about 960 (her floor from the income target stays 480.56).
• Is demand above capacity? Utilization and the waitlist.
• How do new buyers react to the price? Enquiries, how many said yes, and why the others said no.
• What is the season doing? Booked lessons by month, summer against spring.

Then one decision for the next year: hold, change or test. Mixing the questions up, for example reading the summer as a verdict on the rate, is how solo founders talk themselves into a discount they don't need.

How it fits together: once a year, one page: realized price, a price–volume–mix split in sales, costs, checked competitor prices, test results and one decision. The case comes next: the recommendation that started Klasna's September change.