← Pricing Strategy

Module 5: Testing & Optimizing

Guessing at prices is expensive. Testing them is cheap — if you do it right. This module teaches you to run clean price tests, read the one metric that actually matters, and stay calm when the numbers wobble.

Variable you change in any single test

100

Minimum visitors per variant before you judge

Days between pricing reviews — quarterly discipline

Testing prices without chaos

A price test only works when you trust the result — and you only trust it when you change one thing. Call it The One-Thing Test: change exactly one variable, the price, and nothing else. Not the price plus a redesigned page plus new photos. If sales rise, the price must get the credit. Change two things and you learn nothing — just a new page and a shrug. One test, one change, one lesson.

Give every test enough data: run each variant at least two full weeks, or until each variant has seen 100 visitors — whichever takes longer. Two full weeks capture the weekday-weekend rhythm of buying; 100 visitors keep one random buyer from deciding the whole test. A test with 23 visitors is not a test. It is an anecdote wearing a lab coat.

Write everything down before you start: the start date, the end date, the exact change ("$47 to $67, nothing else touched"), and visitors and sales per variant. Memory lies — in a month you will swear the test ran longer and the price was different. A notebook or a spreadsheet, one row per test, filled in as you go — the written record turns a hunch into a system.

Split traffic: half see the old price, half see the new one. If you cannot split traffic — one email list, one landing page — run back to back: two weeks old, two weeks new — and note the data is messier. Either way, no peeking: do not end the test early because day two looks bad. Bad days are exactly why the test runs two weeks.

A clean two-week test — Variant B wins

Variant A — $2912% convert · $3.48 per visitor

Variant B — $3910% convert · $3.90 per visitor

Build brief

What is the first price test you would run? Write your current price, the test price, the start date, and how you will get 100 visitors to each variant. What is the one thing — and only one thing — that changes?

Reading the signals

The metric that matters is revenue per visitor — not conversion rate. Most people watch conversion, the percentage of visitors who buy, and panic when it drops. But conversion is a vanity metric for pricing decisions: it tells you how many people bought, not how much money you made. You cannot pay your costs in conversions.

Run the numbers. At $47: 100 visitors produce 10 sales, which is $470 — or $4.70 per visitor. At $67: 100 visitors produce 8 sales, which is $536 — or $5.36 per visitor. Conversion dropped from 10% to 8%, and revenue went up. The higher price wins.

This is why the one-thing test and revenue per visitor belong together. The test gives you clean numbers; revenue per visitor turns those numbers into a decision. Compute it as total revenue divided by total visitors for each variant, and pick the bigger number. No debate, no gut feeling, no arguing about what the drop in conversion "means." The math decides.

The same math protects you in reverse. Sometimes the lower price genuinely wins: $29 converting at 20% earns $5.80 per visitor, which beats $49 converting at 10% at $4.90 per visitor. Either direction is a fine outcome. The point is never that higher is better — the point is that the feeling is never the decider. Do the math, not the feeling.

Build brief

Pull your last month's numbers: total visitors and total sales revenue. Compute your revenue per visitor. Now imagine raising your price 20% while conversion drops 15% — does your revenue per visitor rise or fall? Show the math.

When to hold your price

Not every dip means cut. Sales sag for a dozen reasons that have nothing to do with your price: traffic quality shifted and a viral post sent you tire-kickers, the calendar turned and your product always dips this season, or the sample is simply too small and forty visitors are making noise. Price is the last thing you touch — because every price change resets everything you have learned.

Run the checks in this order. First, traffic: are visitors arriving from the same sources in the same mix? Second, season: does this dip happen at this time in every cycle? Third, sample: two full weeks, 100 visitors per variant — do you actually have that? Only when all three check out do you even consider the price. The businesses with the strongest pricing are the calmest about it.

There is a second reason to stay calm: weekly price fiddling trains your customers to wait. If the price moves every few days, buyers learn to hold off — it will be cheaper next week. A steady price does the opposite: customers buy now instead of trying to time the market. Stability is a feature of your pricing, not a lack of strategy.

Make calm a discipline with The Calm Rule: revisit pricing quarterly — put it on the calendar — and never fidget weekly. At each review, look at revenue per visitor, the tests you ran, your costs, and what has genuinely improved. Then decide: raise 10% using the 10% Rule from Module 4, hold steady, or design one new test. Then close the spreadsheet for three months.

Before you touch the price

What a dip in sales usually is

3checks first

Traffic quality40%
Season / timing25%
Sample too small20%
Actual price15%

Example for illustration — not real results. Run the lesson's three checks (traffic, season, sample) before considering price.

Build brief

When did you last change — or seriously consider changing — your price, and was it data or a feeling? What date will your next quarterly pricing review be? Write it down and put it on your calendar.

Marcus sells a $97 video course. After one slow week — only 40 visitors — his conversion dips from 8% to 5%. He wants to cut the price to $67 immediately "to get sales moving." What should he do?

Your turn: Design your first price test on paper — your current price, your test price, the one thing that changes (and what stays the same), your start date, your end date (at least two full weeks out), and the metric you will judge by: revenue per visitor.

Your next 24 hours

Open your analytics, pull your last 30 days of visitors and sales revenue, and compute your revenue per visitor. Write it down somewhere permanent. That is your baseline — every future test gets judged against it.

Steal this template

Price test log

Test name: ___
Start date: ___ · End date: ___
The ONE change: $___ to $___ (everything else stays the same)
Visitors — Variant A: ___ · Variant B: ___
Sales — Variant A: ___ · Variant B: ___
Revenue per visitor — A: $___ · B: $___
Winner: ___
Next action: ___

Module 5 cheat sheet

The One-Thing Test: change the price and nothing else, or you learn nothing.
Run each variant at least two full weeks, or 100 visitors — whichever takes longer.
Judge by revenue per visitor (revenue divided by visitors), not conversion rate.
Log every test: dates, exact change, visitors, sales — memory lies.
The Calm Rule: review pricing quarterly on the calendar; never fidget weekly.
When sales dip, check traffic, season, and sample size first — price last.