Reading the Shadow Dashboard

The shadow dashboard shows the results of evaluating real checkout traffic against your staged changes. Here’s how to read it.

Coverage metrics

At the top of the dashboard, you’ll see:

Metric	What it means
Total checkouts	How many real checkouts have been evaluated since you entered testing mode
Unique zone/group combos	How many distinct zone + product group combinations have been seen
Total possible combos	How many zone + group combinations exist in your configuration
Coverage %	(Unique combos seen / Total possible combos) x 100
Rate changes	How many checkouts would have shown different rates under the staged configuration

What coverage percentage means

• 100% — Every zone/group combination has been tested. Maximum confidence.
• 60-99% — Most combinations tested. Good confidence for common scenarios; rare zones might not be covered.
• Below 60% — Many combinations untested. Consider waiting for more traffic or using Test Rates to simulate missing scenarios.

Tip

Even with 100% coverage, verify edge cases using the Test Rates panel. Coverage tells you that each zone/group combo was seen, not that every possible cart configuration within those combos was tested.

Shadow result details

Below the metrics, you’ll see a list of individual checkout evaluations:

Each entry shows:

• Timestamp — when the checkout happened
• Cart summary — destination, number of items, cart total
• Live rates — what the customer actually saw
• Staged rates — what they would have seen with your changes
• Difference — highlighted when live and staged rates don’t match

Entries with differences are marked with a warning icon so you can spot them quickly.

Making the promotion decision

Before promoting, check:

1. Are the rate changes intentional? Every difference between live and staged rates should be something you expected. If you see unexpected changes, investigate before promoting.

2. Is coverage sufficient? If important zones have no traffic yet, use Test Rates to simulate those scenarios manually.

3. Are there zero-rate surprises? Watch for methods that show $0.00 in staged but had a price in live — this might mean a rule is accidentally making something free.

4. Are any methods missing? If the staged rates have fewer methods than live, a Hide rule might be firing unexpectedly.

When you’re confident, click “Promote”. The staged changes are applied atomically — all at once, in a single transaction. A new version is created with the label “Promoted staged changeset”.

If something goes wrong after promoting, you can roll back to the version before the promotion using Version History.