# Fair and lawful price tests > What to vary, who takes part, what to disclose, and Dana's small sequential test. - Course: [Pricing: Set, Package and Raise Your Prices](https://productovi.com/courses/pricing.md) - Module: How do we test and review prices? - Lesson 29/32, ~23 min - Source: https://productovi.com/courses/pricing/fair-price-tests ## Test before you change **Alex Tutor:** Over this course we've set, packaged, presented, discounted and changed Klasna's prices. This last module is about learning what a price does before you commit to it, and reviewing prices every year. **Alex Tutor:** It's January 2027. Nazar wants to know whether families would pay more than 620 UAH a lesson for the English 16-pack, before he plans September. **Nazar:** Let's test 670 against 620 in February and March. Half the visitors see one price, half the other. **Leo:** Why random? Show the higher price to people on expensive phones and to families who come back three times. They clearly want it more. **Alex Tutor:** That's no longer a test, Leo; it's a personalized price, set from someone's data. As of October 2026, EU rules require telling consumers when a price was personalized by automated decision-making. And a family that finds out feels cheated, whatever the law says. **Alex Tutor:** A fair test varies the price by chance or by time, not by who the person seems to be. Each visitor keeps the same price on every visit, and nobody is charged more for being who they are. > **Exercise:** Which price-test designs are fair and lawful for Klasna? Select all that apply. ## The house rules **Hanna:** Who's in it? I don't want a family that has paid 620 for two years suddenly seeing 670. **Alex Tutor:** Only new visitors in Ukraine, paying in hryvnia. Existing families see the list price. Families abroad stay out: their EUR prices are another question, and EU consumer rules apply to them directly. **Alex Tutor:** And a safety net: if the test ends at the lower price, everyone who bought at the higher one gets the difference back as lessons. Social discounts are never part of a test. **Alex Tutor:** So who exactly sees a test price? > **Exercise:** Who should take part in Klasna's February–March test of the English 16-pack price? **Nazar:** And how long do we run it? **Alex Tutor:** Decide before you start: the main measure, the length and the sample. Here the measure is revenue and contribution per visitor plus repeat purchases, not conversion alone. How to size a test is covered in the course Product Analytics: From Data to Decisions. **Alex Tutor:** Which number should decide the test? > **Exercise:** Which main measure should Klasna write down before its price test starts? > **Exercise:** **True or false:** if a price test ends at the lower price, it is fair to keep the extra money from families who bought at the higher test price. ## Dana's small test **Alex Tutor:** Now a different business again: Dana's practice. She gets two or three enquiries a month. Splitting them in half would tell her nothing. **Alex Tutor:** Could Dana split her enquiries the way Klasna splits visitors? > **Exercise:** **True or false:** with two or three enquiries a month, Dana gets a reliable answer within a month by splitting them at random between 600 and 750. **Dana:** So I tried it one after the other. From 15 March to 30 April I quoted 750 to everyone new, and my price post said 750 too. My learners and my waitlist kept 600. **Dana:** Before that, 11 of 14 enquiries started at 600. In the test, four of six said yes to 750. One said no because of the price, one because of the schedule. > **Exercise:** In Dana's test, 4 of 6 new enquiries accepted 750 UAH. What share is that? In %, to two decimals. **Sofia:** 66.67% against 78.57%. So the higher price lost some people? **Alex Tutor:** Six enquiries can't measure a rate. One more yes or no moves it by 17 points. And spring is her busiest time, so the two periods differ in more than price. **Alex Tutor:** What a small test can do is look for a wall. Dana wrote her rule before starting: keep 750 if at least half accept. Four of six did, so there's no wall at 750. > **Exercise:** From 15 March to 30 April 2027, in her busiest season, Dana quoted 750 UAH to every new enquiry. Before starting she wrote a rule: keep 750 if at least half of new enquiries accept it. Four of six did; before the test, 11 of 14 enquiries had started at 600. What can Dana honestly conclude from her test? > **Exercise:** Put the steps of a fair price test in order. ## Your turn: Klasna's test rules **Nazar:** Write me the rules for the February test, so I can show Hanna and Leo before we start. > **Exercise:** Write Nazar the rules for Klasna's February–March test of the English 16-pack (620 vs 670), in 4–6 sentences: who takes part, how the price is assigned, the safeguards and how the result will be judged. > **Key ideas** > > - **Vary the price by chance or by time**, never by who the buyer seems to be. Personalized prices set by automated decision-making must be disclosed to EU consumers (as of October 2026). > - **House rules:** new visitors only, the same price on every visit, existing customers at the list price, a refund of the difference if the test ends lower, no social discounts in tests. > - **Decide before you start:** the main measure (revenue and contribution per visitor, retention), the length and the sample. > - **Small businesses test in sequence**: one price for everyone new for a period, written up in advance. > - **A tiny test finds walls, not rates:** Dana's 4 of 6 at 750 (66.67%) says there's no wall, not that the rate fell from 78.57%. **Alex Tutor:** Next we read Klasna's February–March test, and one of its arms has a surprise in it. - Previous lesson: [Case: the September 2027 price change](https://productovi.com/courses/pricing/september-price-change-case.md) - Next lesson: [Reading a price test](https://productovi.com/courses/pricing/reading-a-price-test.md) --- Productovi Learning. Free, no sign-up needed to read.