Before you start
- Messages on (active mode) for scenarios 1 to 7, 9 and 10. Scenario 8 turns them off.
- Use a private or incognito window for each fresh shopper, so you are a new visitor each time.
- “Wait” means real waiting. The app reads stillness; moving the mouse or scrolling resets the clock.
- Your facts decide what can be said. The page reads your sheet and tells you, per scenario, whether there is a fact to state.
The scenarios
1
1. Stuck on shipping
Open the store in a private window. Add one item and open the cart. Stop for about 45 seconds.Expect: one sentence about free shipping, naming something that gets the cart over the line. If your store has no free shipping line, nothing, because there is nothing true to say.Proves: the card, the timing, and that it uses your facts.
2
2. A code that doesn't work
In the cart, type a code that does not exist, apply it, then try a second one. Wait about 15 seconds.Expect: with offers on, the card says that code is not valid here and names a real one. With offers off (the default), it says the code is not valid here and gives a fact instead.Proves: declined codes are noticed and answered without giving away margin by default.
3
3. Can't decide on size
Open a product with sizes. Switch the size three or four times without adding to the cart. Wait about 45 seconds.Expect: the card gives that product’s note. If no product has a note yet, nothing, and the page says so. On a test store, Add a sample note writes “Runs small. Most people take one size up.” on your first sized product.Proves: silence when there is no fact, a fact when there is.
4
4. A normal shopper
Browse a few products at a normal pace, add one, check out.Expect: nothing is shown. The list on the right stays quiet for this visit.Proves: happy shoppers are left alone. This visit lands in neither comparison group.
5
5. Stuck, then bought
Do scenario 1 and let the card appear. Check out (on a development store, with Shopify’s test gateway).Expect: within a minute the order shows on Money as credited to that moment, with its order reference. On a test store, press Run the meter: one usage record for 5% of it appears under Billing.Proves: the credit, and the charge, keyed to the order.
6
6. Bought, then stuck again
Right after scenario 5, in the same window, add another item, open the cart and wait 45 seconds.Expect: nothing. Reasons shows the visit as “bought in the last day”.Proves: no one who just paid hears from the app for 24 hours.
7
7. Held back
Repeat scenario 1 in ten fresh private windows.Expect: about one in ten hears nothing and shows as “Held back to compare”. The rest hear the sentence. (A new store holds back one in five until 30 have been held.)Proves: the comparison group is real and random.
8
8. Watch mode
Turn Messages off in Settings. Do scenario 1 twice.Expect: nothing is shown, but both visits appear on the right as “Left alone, nothing shown”, with the sentence that would have been said.Proves: watch mode records everything and shows nothing.
9
9. The cap
In Settings, change the cap to $25 and approve it. Do scenario 5 enough times that 5% of the orders passes $25, pressing Run the meter after each. Then raise the cap back.Expect: the meter reports capped, Home shows the pause message with how many shoppers were missed since, Messages shows off. Raising the cap turns them back on.Proves: the pause, the message, the resume.
10
10. Another country
Switch the storefront’s country selector to a country you have no shipping zone for, or to another currency, and do scenario 1.Expect: nothing about shipping is claimed.Proves: no sentence a shopper could prove wrong. This scenario is in progress in NextIntent’s own test bank; the behaviour is as described, the scripted check is still being added.