Pause or stop?
Both are available on a running test. Pause removes the test from your store — visitors see normal prices again — but keeps it resumable. Stop ends it for good.
Because a paused test is no longer active on the store, prices revert while it is paused and no data is collected. Resuming puts it back.
Assignment is deterministic per visitor and test, so a visitor who returns after a resume lands in the same group as before. What you can lose is visitors whose qualifying session has since ended — if the test has audience filters, someone who no longer matches will not re-enrol.
Use pause sparingly. A test with a gap in the middle mixes two periods of traffic, and any seasonality or campaign activity in between quietly contaminates the comparison. If something is broken, pause it; if you are just unsure, let it run.
When to end
End it when you have an answer, not when you have a number you like:
Enough orders in each arm that one more would not change the conclusion
At least two to four weeks elapsed
The result has held steady for several days rather than swinging
End it immediately if shoppers are being charged the wrong price, checkout is broken for tested products, or prices display inconsistently enough to confuse customers.
How to end it
Open the test and click Stop. What happens:
The test is no longer published, so it drops out of the store's active price tests
Test prices stop being displayed — everyone sees your normal Shopify price
Assignments already stored in visitors' browsers are pruned on their next visit, so nobody stays stuck on a test price
Reporting stops collecting; results remain available
Your Shopify product prices are untouched throughout, so there is nothing to restore.
A stopped test cannot be edited or restarted. To run it again, create a new test — which is the honest option anyway, since a restarted test would blend two separate periods of traffic into one result.
Reading the result
Look per product, not just at the overall total. A test across four products can be a win on two and a loss on two, and the aggregate hides that. The Product report tab exists for exactly this.
For each product you are looking at one of three outcomes:
A clear winner. One price produced meaningfully more revenue. A higher price can win on revenue while losing on conversion — that is still a win if margin is what you are optimising for.
No measurable difference. Genuinely useful: the market is not sensitive to price in that range. If the higher price is one of them, take it — same demand, more margin.
A clear loss. Conversion or revenue dropped. Keep the original price, and you have learned where the ceiling is.
Applying the result
There is no automatic "apply winner". When you have decided, change the price in Shopify admin yourself. Before you do:
Check the result per product, and apply per product
Confirm the revenue effect is real, not an artefact of one large order
Check margin, not just revenue — a price that wins on revenue can lose on profit if it shifted the mix
A note on the net sales figure
Net sales in the product-level report is calculated from tracked storefront events rather than from Shopify's order records, so it will not reconcile with the revenue figures in your Shopify admin.
It follows Shopify's own vocabulary: gross sales are before discounts, net sales after them. The figure here is net of discounts — the price used is what the shopper actually paid per unit — but it excludes tax and shipping, and unlike Shopify's own net sales it is never reduced by refunds or cancellations, because it is recorded at the moment of purchase.
That is fine for what the report is for: comparing one arm against another, both measured the same way. It is not the number to quote as store revenue.
After ending
Record what you learned alongside the hypothesis
Record "no difference" results too — they stop someone re-running the same test in six months
If you found a winner, consider whether the same question applies to similar products