**Amazon Ads → Keywords** lists every keyword you bid on, across campaigns and ad groups, with its **match type** beside it.

Match type is the column that changes how you read everything else on the row — and it is the one people skim past.

[Watch the walkthrough](https://kb.trackiq.com/kb-assets/kbvideo-keywords-explainer-a7fbbf53.mp4)

*Prefer to watch? Seventy-two seconds on match types, and why you act on exact but investigate broad.*

## **The screen**

![The Keywords screen](/kb-assets/shot-adv-keywords-view-ce68f0.png)

Same layout as the rest of Amazon Ads: six headline tiles with sparklines, Performance Trends with **Day / Week / Month**, **vs previous**, and **Values / % change**, then the table.

The table identifies each keyword by where it lives:

| Column | |
|---|---|
| **Campaign** | which campaign |
| **Ad group** | which ad group inside it |
| **Keyword** | the term you bid on |
| **Status** | enabled, paused, archived |
| **Type** | the ad type — SP, SB |
| **Match type** | broad, phrase or exact |

**A keyword is not unique.** The same word can appear several times at different match types, in different ad groups, at different bids — and each row is a separate thing you are buying. Sorting by keyword rather than by spend makes that obvious, and is worth doing once.

## **What match type actually controls**

Match type governs how far Amazon may stray from your keyword when deciding whether to show your ad.

| Match type | Amazon shows your ad for |
|---|---|
| **Exact** | that term, and close variants of it |
| **Phrase** | that term as part of a longer search |
| **Broad** | searches it judges related |

Reading them as a range of *control against reach* is the useful framing: exact is precise and limited, broad is wide and speculative.

## **Why that changes how you read a row**

**A broad keyword's performance is not really about the keyword.** It is about the pile of searches Amazon matched to it. A broad keyword with poor ACoS may contain excellent searches and terrible ones averaged together.

So a bad broad keyword is a **prompt to look at search terms**, not a prompt to cut the bid. The bid change applies to everything it matched, good and bad. See [**Search Terms and Aggregated Search Terms**](https://kb.trackiq.com/hc/knowledge-base/articles/search-terms-and-aggregated-search-terms).

**An exact keyword's performance is about that keyword**, so acting on it directly makes sense.

That is the practical rule: **act on exact, investigate broad.**

## **What good structure looks like**

Most well-run accounts run the same loop:

1. **Broad and phrase keywords discover** which searches convert.
2. Winning searches are **promoted to exact keywords**, usually in their own ad group where the bid can be set deliberately.
3. Losing searches become **negatives**, so the broad keyword stops paying for them.

The Keywords screen is where step 2 lands and step 3 gets checked. The Search Terms screen is where steps 1 and 3 start.

## **Reading the table**

**Sort by spend, not by ACoS.** A keyword at 300% ACoS on $4 is not the problem; one at 60% on $4,000 is.

**Watch the click count before believing a rate.** ACoS on six clicks is not a measurement. This is the same threshold logic the rule engine uses — see [**Guardrails to Set Before You Automate Anything**](https://kb.trackiq.com/hc/knowledge-base/articles/guardrails-before-you-automate).

**Check the state filter.** Enabled-only is the default, so paused keywords with history are hidden.

**Note the ad type.** A Sponsored Brands keyword is judged on a 14-day window against Sponsored Products' 7. Do not rank them against each other on efficiency.

## **Acting on what you find**

Manual bid changes here are fine for one-off decisions. For anything recurring, the rule engine does the same judgement continuously:

- **Hit my Target ACOS** scales every bid by distance from your target.
- **Pause dead keywords** stops targets that have spent enough clicks to prove they will not convert.
- **Rescue impression-starved winners** raises bids on keywords that convert well but rarely show.

See [**Targeting Rules: Adjusting Bids and Pausing Targets**](https://kb.trackiq.com/hc/knowledge-base/articles/targeting-rules-bids-and-pausing).

## **Related**

- [**Search Terms and Aggregated Search Terms**](https://kb.trackiq.com/hc/knowledge-base/articles/search-terms-and-aggregated-search-terms) — what customers actually typed.
- [**SQP & Keywords: Finding Queries You Convert On But Don't Bid On**](https://kb.trackiq.com/hc/knowledge-base/articles/sqp-and-keywords-finding-unbid-queries) — demand you are not buying at all.
- [**Understanding ROAS and Key Amazon Advertising Metrics**](https://kb.trackiq.com/hc/knowledge-base/articles/understanding-roas-and-key-amazon-advertising-metrics) — what the tiles mean.
- [**Amazon PPC: Core Terms and Metrics**](https://kb.trackiq.com/hc/knowledge-base/articles/amazon-ppc-terms-and-metrics) — match types defined, with the rule that matters most.
- [**Create a Sponsored Products Campaign**](https://kb.trackiq.com/hc/knowledge-base/articles/create-a-sponsored-products-campaign) — choosing them at campaign creation.
- [**Bid Optimisation: Six Rules**](https://kb.trackiq.com/hc/knowledge-base/articles/bid-optimisation-six-rules) — the rules applied to these targets each week.
- [**Outlier and Trending Keywords**](https://kb.trackiq.com/hc/knowledge-base/articles/outlier-and-trending-keywords) — two sources of keywords worth adding.
- [**Grouping and Prioritising Keywords**](https://kb.trackiq.com/hc/knowledge-base/articles/grouping-and-prioritising-keywords) — organising them into launch groups.