Each of these follows the same shape: a decision you would otherwise make on instinct, and a way to make it on evidence instead.
Raise prices without guessing
The problem. You suspect a product is underpriced, but raising the price feels like a gamble. If conversion drops more than the margin gains, you lose — and by the time you notice, you have lost a month.
How a test helps. Half your visitors see the higher price, half see the current one. You find out what actually happens to conversion and to revenue per visitor, on a slice of traffic rather than all of it.
Example. A moisturiser sells at 1,200. You test 1,400 on half your traffic. Conversion falls 8%, but revenue per visitor rises 9% — the increase more than pays for the lost conversions.
Respond to a cost increase
The problem. A supplier raises costs or a tariff lands, and margin has to come from somewhere. The instinct is an across-the-board increase, which risks damaging your best sellers.
How a test helps. Test the increase on the products you are least sure about before applying it everywhere. You will often find some products absorb it with no measurable effect and others do not.
Example. Costs rise 12%. Rather than adding 12% to everything, you test it on your top five products first and discover two of them lose conversion badly — those get a smaller increase.
Price a new product
The problem. A new product has no history, so there is nothing to base a price on. Whatever you pick becomes the anchor customers judge it by.
How a test helps. Launch with two candidate prices at once and let early traffic tell you which performs. You get an answer in the first weeks rather than after a quarter of selling at the wrong price.
Check whether you are priced right
The problem. Prices set eighteen months ago are still in place because nobody has revisited them. They may be leaving money on the table, or quietly suppressing conversion.
How a test helps. Test a small move in each direction on a handful of products. Even a result of "no measurable difference" is useful — it tells you there is room to move on price without losing sales.
Test a psychological price point
The problem. Is 999 meaningfully better than 1,000? Everyone has an opinion, nobody has evidence for your store.
How a test helps. The difference is small enough that only real traffic can settle it. Because the change is tiny, this needs more traffic and patience than a large move.
Test the discount, not just the price
The problem. You want to know whether a bigger struck-through price sells more, independently of what the shopper actually pays.
How a test helps. An arm can adjust the compare-at price as well as the selling price, so you can present the same price at a different apparent saving and see which converts better.
If you set no compare-at adjustment, Cooee handles the labels for you: when the test price is lower than the original, the original is shown struck through with a savings label.
When a price test is not the right tool
You want everyone to see the new price. That is a price change, not a test. Change it in Shopify.
You need an answer this week on a low-traffic product. A product with a handful of orders a week will not produce a readable result.
You want to run a sale. Use a discount. Price tests are for learning, not promoting.
The product is already in another active price test. Only one adjustment is ever applied, and both tests become unreadable.