Free Shipping Threshold Test: Find the Number That Pays

I picked my first free shipping threshold in about four seconds. A competitor used $50. So did I.

That's how most founders do it, and it's why so many thresholds quietly lose money. Here's the short answer. A free shipping threshold test splits your traffic between two or three thresholds, keeps each buyer on one number from the first banner through checkout, and crowns the winner on margin per visitor after you pay the carrier. Not on average order value. Not on conversion rate.

The 4-step version:

1. Pull 90 days of orders and find where order values cluster.
2. Test your current threshold against one or two numbers near those clusters.
3. Change every surface that prints the threshold, in every group.
4. Run two to four weeks, then pick the group that keeps the most margin per visitor.

The right threshold isn't the one that grows the cart. It's the one that leaves the most money in your account after the shipping label is paid.

Why does a gut-feel threshold cost you money?

Because a threshold can fail in two opposite directions, and both feel fine on a dashboard.

Set it too low and you're paying for shipping on orders that were already happening. Intelligems' own threshold test guide puts it plainly: too low and "you subsidize shipping on orders that would have converted anyway." Set it too high and buyers stop trying. They look at the gap, decide it isn't worth it, and leave.

Both mistakes hide well. The low threshold shows a lovely conversion rate. The high threshold shows a lovely average order value. Neither number tells you whether you made money.

And free shipping does matter to buyers. Baymard Institute's cart abandonment research lists extra costs (shipping, tax, fees) as a reason behind 40% of abandoned carts. That's exactly why a threshold is worth testing instead of guessing. It moves real behavior, in both directions.

What should you actually test?

Start with your own orders, not a competitor's banner.

Intelligems suggests pulling 90 days of orders, sorting them by value and looking for where they bunch up. Their threshold article gives two red flags worth taping to your monitor: if fewer than 30% of orders land within $10 of your threshold, it may be too high. If more than 50% hit it exactly, you're likely leaving money on the table.

Their test guide puts the usual starting zone between your current average order value and about 20% above it. That's the range where buyers will add one more item instead of walking.

This is also where AI saves you an afternoon today. Export your orders as a CSV from Shopify, hand the file to Claude or ChatGPT, and ask for a histogram of order values in $5 buckets plus the share of orders within $10 of each candidate threshold. You get the cluster map in minutes, and you can check every number against the raw file.

Then pick two or three arms. Your current number is always the control.

How do you run a free shipping threshold test without breaking checkout?

This is the hard part, and it's where most DIY tests die.

The threshold doesn't live in one place. Shoplift's threshold testing write-up counts 8 to 15 surfaces where the number shows up: the announcement bar, product pages, the cart progress bar, emails, the shipping policy page. If the banner says $75 and checkout charges shipping under $100, you didn't run a test. You ran a trust leak.

So use a tool that can change the actual shipping rate per visitor, all the way to checkout. Two details from Intelligems' shipping testing FAQ are worth knowing before launch:

• You can lower a threshold with a Free Shipping experience, but raising one requires a Shipping Rates experience.
• Thresholds are calculated in your store's default currency, and on the post-discount subtotal (a change they made in early 2026). Check which subtotal your own setup uses, because a 20% off code can push a buyer under the line.

Then QA the whole path in every arm. Add to cart, watch the progress bar, open checkout, confirm the rate. If you've never run a clean split before, the store owner's guide to ecommerce A/B testing covers the sample size side.

Which metric crowns the winner?

Margin per visitor after shipping. One number. Write it down before launch so nobody can move the goalposts when the results come in.

Intelligems calls profit per visitor "the most important metric" in a threshold test. Shoplift goes further and defines it as contribution margin per visitor: revenue minus product cost, absorbed shipping and fulfillment, divided by every visitor in the group. I agree with both. Conversion rate and average order value are the diagnostics. They tell you why a group won. They don't tell you who won.

Now watch what happens on a hypothetical store. Run the math on a loose-leaf tea brand like this: $28 tins, a 65% gross margin on product, and a $9 average shipping label. Below the threshold, the buyer pays the $9. Above it, the store eats it. Three groups of 20,000 visitors over four weeks:

Threshold	Conversion rate	Average order value	Orders shipped free	Revenue per visitor	Margin per visitor after shipping
$50	3.0%	$61	80%	$1.83	$0.97
$75	2.8%	$74	60%	$2.07	$1.20
$100	2.3%	$82	30%	$1.89	$1.16

Here's the math on each, so you can check me:

• $50: 600 orders x $61 = $36,600 revenue. Product margin $23,790 (65%). Shipping absorbed: 480 free orders x $9 = $4,320. That leaves $19,470, or $0.97 per visitor.
• $75: 560 orders x $74 = $41,440 revenue. Product margin $26,936. Shipping absorbed: 336 x $9 = $3,024. That leaves $23,912, or $1.20 per visitor.
• $100: 460 orders x $82 = $37,720 revenue. Product margin $24,518. Shipping absorbed: 138 x $9 = $1,242. That leaves $23,276, or $1.16 per visitor.

$50 threshold $61

$75 threshold $74

$100 threshold $82

20,000 visitors per group. The $100 threshold wins this chart, and it's the wrong chart to judge by.

$50 threshold $0.97

$75 threshold $1.20

$100 threshold $1.16

(Revenue x 65% product margin, minus $9 for every order shipped free) divided by 20,000 visitors.

Look at what each dashboard would tell you. The conversion rate crowd picks $50. The order value crowd picks $100. The margin picks $75.

And here's the math that matters. On 60,000 monthly visitors, the $75 threshold keeps $71,736 in margin after shipping. The $50 threshold keeps $58,410. That's $13,326 a month from changing one number in your shipping settings.

A higher threshold always grows the cart. It grows it partly by sending away the buyers who couldn't reach it.

When does the test lie to you?

When you read a small gap like a big one.

In that table, $75 beats $100 by 4 cents per visitor. With about 500 orders per group, that gap may be noise. What the test does prove: $50 is clearly bleeding margin, and anything from $75 to $100 is safer. I'd roll out $75 and run a second test against $90 next month.

Three other ways a threshold test fools you:

• A sale in the middle. A sitewide promo changes cart sizes for everyone. Shoplift recommends steering clear of major promotions, and I'd add holidays too.
• Returns. Buyers who pad a cart to hit a threshold sometimes send the padding back. Shoplift treats return rate as a guardrail. Check it 30 days after the test ends.
• Calling it early. I've watched founders stop a test on day five because the higher threshold "lost." Set the end date first. Intelligems expects results in 2 to 4 weeks, depending on traffic.

Why does the product page decide how high you can go?

Because a threshold asks the buyer one silent question: is it worth adding another item just to save $9?

A spec sheet can't answer that. A page that makes the second tin feel obvious (a gift, a backup, the flavor that pairs with the first) turns the threshold into a nudge instead of a wall. That's the same job a free shipping bar does in the cart, and it's why free shipping thresholds raise order value on some stores and do nothing on others. A mix and match bundle priced to clear the threshold can do the same job in one click.

Look at what happened on the Copper Bamboo Sheets page for one of our client stores. Real client numbers, not typical results, and not a promise of what your store will do. Before: conversion rate 1.0%, average order value $125, revenue per visitor $1.25. After the page rebuild: conversion rate 4.3%, average order value $221, revenue per visitor $9.50. On 10,000 visitors that's $12,500 against $95,000. The full before and after numbers are on our results page.

Order value went up and conversion rate went up at the same time. That's what a page that sells does. It lets a higher threshold work without scaring buyers off.

What to do next

Export the last 90 days of orders tonight. Find where they cluster, write down your margin and your average shipping label cost, and pick one threshold to test against your current one. Then write "margin per visitor after shipping" at the top of the test plan, before you launch, so nobody crowns the wrong winner.

Pricing tests follow the same rules, so read how to run Shopify price testing next. And if your team still argues over which number matters, settle it with revenue per visitor versus conversion rate, then work out profit per product so you know which margins you're protecting.

---

Book Your Profit Audit

A free shipping threshold only climbs as high as your product page can carry the buyer. Book a profit audit and we'll show you where your margin per visitor is leaking, then how to rebuild a high-converting product sales page in less than 15 minutes.

Book Your Profit Audit →

Or go here to check it out → revenueflows.ai

P.S. The $50 threshold won on conversion rate and lost $13,326 a month in margin. Count what you keep, not what you ring up.