# Understand Labelizer buckets and results

> Labelizer groups products into performance buckets so you can quickly see which products are above, near or below target.

[Labelizer](/docs/labelizer/labelizer) groups products into performance buckets so you can quickly
see which products are above, near or below the target you set. Each Labelizer project has its
own target, settings, runs and results.

## What buckets mean

Labelizer compares each product against the target you configured during setup. Typical bucket
groups are:

| Bucket      | Meaning                                                     |
| ----------- | ----------------------------------------------------------- |
| Over-index  | Products performing clearly above the target                |
| Index       | Products performing around the target                       |
| Near-index  | Products close to the target but not clearly above it       |
| Under-index | Products performing below the target                        |
| No index    | Products without enough usable data for a performance label |

The exact result depends on your selected metric, target, date range, conversion lag and filters. If
you want the rules themselves, including the click thresholds and how Index margin and Click
multiplier feed into them, see
[How Labelizer decides each bucket](/docs/labelizer/how-labelizer-decides-each-bucket).

## ROAS and CPA targets

If your project uses ROAS, higher values are usually better. If your project uses CPA, lower
values are usually better. Labelizer uses the selected target to decide whether a product is
performing above, near or below expectation.

## Why products can end up in No index

A product may be placed in No index when Labelizer cannot calculate a reliable bucket. Common
reasons are:

- The product has too little traffic.
- The product has no conversions in the selected date range.
- The product does not match the feed or campaign filters.
- The product was not active during the selected period.
- Conversion lag means recent conversions are not counted yet.

No index does not always mean something is wrong. It often means there is
[not enough data for a confident label](/docs/labelizer/why-labelizer-did-not-update-custom-labels).

## Review run results

Open the Labelizer project and review the latest run. Check:

- When the run started and finished.
- Whether the run succeeded or failed.
- How many products were processed.
- How products were distributed across buckets.
- Whether the results match the account and filters you expected.

See [Run Labelizer and review results](/docs/labelizer/run-labelizer-and-review-results).

## What to do with the results

Use the buckets to decide which products need attention. For example:

- Over-index products may deserve more budget or visibility.
- Under-index products may need bid, feed, price or landing page review.
- No index products may need more data before you act.

Do not treat the buckets as a replacement for campaign context. Use them as a starting point for
product-level decisions.