Find Your Pricing Cliff: The 24-Hour Price Test Most DTC Brands Never Run

When did you last actually test your price? Not nudge it for a sale. Not raise it because costs went up. I mean deliberately put two prices in front of two halves of your traffic and let the data tell you which one your customer prefers. If the honest answer is "never", you are almost certainly leaving profit on the table right now, today, and it's the cheapest profit you'll ever find.

Here's why this one nags at me. Price is the only lever you have that costs nothing to pull. A creative test costs you ad spend. A new product costs you stock. Changing your price costs you a few minutes in Shopify. And yet it's the thing most brands touch the least, usually defending a number they picked once, years ago, mostly on vibes.

The advice everyone gives you is half right

The standard wisdom is "just raise your prices". And sometimes that's exactly right. There are brands sitting on real inelasticity, where you can lift the price and barely anyone flinches, and the whole increase drops straight to the bottom line.

But "always raise" is lazy, and occasionally it's the worst thing you could do.

Here's the part that gets missed. Going down can win too. Drop a price and yes, your margin per unit shrinks, but if enough new buyers come through the door, you can end up more profitable in total and buy yourself a far better-performing ad account at the same time. That second effect is the one almost nobody prices in.

Let me give you a worked example. Imagine a brand selling a premium accessory at $240. Decent product, but it's the most a lot of their customers have ever spent on that kind of thing. They test dropping it to $205. Not a discount, not a sale, just a lower everyday price.

Conversion rate jumps sharply, say up by half again, because that new number sits the right side of a line in people's heads. Margin per unit is thinner, of course. But far more people buy, so total profit per visitor goes up, not down. And here's the bit that compounds: because more of those visitors now convert, the ad account suddenly tolerates a higher cost per click and a higher CAC. You can spend more to acquire a customer because each customer is easier to win. Cheaper price, more adoption, and more room to scale on Meta, all at once. The opposite of what "just raise prices" would've told you to do.

I'm not saying down always wins. I'm saying you don't know which way is up until you test it. My old line on this: we don't know what's best, that's why we test.

What you're actually hunting for: the cliff

The thing you're looking for has a shape, and it helps to picture it.

Across most products there's a cliff. A price point where demand holds up nicely, and then a little higher, it drops off a ledge. Below the cliff, small price changes barely move volume. Cross it, and volume falls off a wall. Your job is to find where that edge sits for your product, because the sweetest price is usually parked right up against it.

The cliff lives in a different spot for every brand, and you genuinely cannot guess it from your chair. You've got too much price memory. You've stared at "$29" for so long that $29 feels like a law of physics, when $39 might sell just as well and $24 might quietly triple your volume. The number in your head is the one thing you can't trust here.

Two things shift where your cliff sits, and they're worth knowing before you test.

First, how many units land in a typical order. If most people buy a single hero product, testing is clean, because the price they're judging is the price of that one thing. If a typical basket has eight different items in it, shoppers anchor on the total at checkout far more than on any single price, and your test gets muddier. Test the simple case first.

Second, the cliff moves over time. The price that felt sharp last January isn't automatically sharp now. Costs drift, competitors reprice, and what counts as "expensive" in your category keeps shifting under you. A cliff you found a year ago is worth re-checking.

How to run the test in 24 hours

1. Pick one product to start. Ideally a hero item that people usually buy on its own, so you're measuring a clean signal and not a tangle of basket effects. One product, one variable.
2. Choose your two prices. Your current price against one challenger. Make the gap meaningful, not timid. A two-dollar nudge tells you almost nothing. Move it enough to actually probe for the cliff, up or down, and pick the direction you're most curious or most nervous about.
3. Split your traffic, not your timeline. Show price A to half your visitors and price B to the other half, over the same window. Running price A this week and price B next week isn't a test, it's two different weeks with different weather, different ads, and different moods. A/B testing tools that sit on top of Shopify will hold the price consistent per visitor and split them cleanly for you.
4. Let it run long enough to mean something. "24-hour" is about how fast you can launch it, not always how fast you read it. On a high-traffic store a day might give you a clear answer. On a quieter one, give it long enough to gather a real number of orders per side before you trust it. Reading a winner off six sales is just noise wearing a costume.
5. Read the right number. This is where most people trip. Do not judge it on conversion rate. The lower price will almost always "win" on conversion rate, and that tells you nothing on its own. The number that matters is profit per visitor, or at the very least revenue per visitor. Take the margin per order, multiply by how many orders each price produced, and divide by the visitors who saw it. The price that makes you the most money per person through the door wins, full stop.

The traps that catch people

A couple of things to watch, because they're the ones that bite.

The obvious one is customers comparing notes. If two people see different prices on the same product and talk, that's an awkward conversation. In practice, with a clean split over a short window it's rarely an issue, but be sensible: test on cold traffic rather than your loyal repeat buyers, and don't run a wide gap on a tight-knit community product.

The subtler trap is reading margin wrong. If your test price involves any discounting or a different shipping setup, make sure you're measuring true profit per order and not just topline revenue. It's entirely possible to "win" on revenue per visitor while quietly making less money, because the thing that lifted conversion also chewed your margin. Always take it down to profit.

And don't over-read a single test. One result on one product is a data point, not a doctrine. The brands that get real money out of this treat it as an ongoing habit, re-testing a few times a year, not a one-off they did once and filed away.

So, what's your number actually worth?

Here's the thing that should bug you a little after reading this. There's a price for your hero product that makes you more money than the one you're charging right now. It almost certainly exists. It might be higher. It might, against every instinct you have, be lower. The only reason you don't know it yet is that you've never put two numbers side by side and let your customers vote.

That's a strange thing to leave untouched, when the test costs you nothing but a few minutes and the patience to read it properly. So the question I'd genuinely sit with is this one: if you ran that test this week, are you quietly confident the price you've got today would win? And if you're not sure, isn't that exactly the reason to find out?