Source: https://docs.amadalis.ai/optimization/slots-and-measurement/

# Slots, performance and measurement

Use **Optimization** > **Personalization** in the site admin to manage where targeted content can appear and how its results are measured. Use **Performance** to compare outcomes for personalised slots and held back visitors.

Checked against the product’s code on 7 November 2026.

Under **Optimization** > **Personalization**, **Slots**, **Performance** and **Measurement** are separate views of the experience.

## Configure a slot

1.  Open **Slots**.
2.  Search for the intended [slot](https://docs.amadalis.ai/reference/glossary/#slot).
3.  Open the slot.
4.  Inspect **What this slot shows**.
5.  Set **Pieces shown at once** (one by default).

If a campaign has pinned a piece to the slot, the slot shows it and the rules below take no part; pins are changed on the campaign.

The optional rules can consider journey stage, freshness, repeated exposure, variety and preferences for promoting products. Their switches are off in the default rule set. Review a rule’s actual bounds before enabling it. A rule can change ranking; an adjustment that promotes products is not merely a label on an otherwise fixed order.

1.  Choose **Save**.
2.  Read the consequence.
3.  Approve the prepared revision if it matches your intent.
4.  Check the receipt and its offered Undo.

An eligible [fallback](https://docs.amadalis.ai/reference/glossary/#fallback) still matters when no targeted item can be shown.

## Read performance

1.  Open **Performance**.
2.  Check the selected scope.
3.  Check the goal.
4.  Read the population.
5.  Read the window and missing data state.

Performance compares recorded targeted and held back outcomes. A recorded difference estimates performance under those conditions; it does not automatically prove that targeting caused an improvement for every visitor.

## Review measurement settings

**Measurement** holds the platform holdout and [attribution](https://docs.amadalis.ai/reference/glossary/#attribution) settings. A visitor held out of personalization is different from an item held back for one comparison. Setting a new holdout salt, the value used to assign visitors to comparison groups, puts every visitor back into a randomly drawn group and restarts the holdout report, so results before and after the change cannot be compared. The card says so before you approve.

Use **Change history** for saved revisions and supported reversals. A rollback changes policy going forward; it does not erase events already recorded. Use [targeted content](https://docs.amadalis.ai/optimization/audiences-and-targeted-content/) and [recommendations](https://docs.amadalis.ai/optimization/recommendations/) for the pieces that fill slots.