Synonyms from the category — Amazon's own taxonomy as a keyword source.

Of all the synonym sources, the category is the cheapest and the most overlooked. Amazon's own taxonomy already encodes the vocabulary it considers related to your product — the work is reading that taxonomy carefully instead of jumping straight to external keyword tools.

Why the category is a privileged source

Amazon's browse-node tree is curated by Amazon's own catalogue team. The node names, the filters that appear on the SERP, and the rails of related products are all explicit signals about what Amazon thinks belongs together. If a synonym appears as a filter on the SERP for your category, Amazon is telling you that shoppers in that category use that word and that the algorithm indexes against it.

The four extraction surfaces inside one category

1. Browse-node breadcrumbs. "Home & Kitchen > Kitchen & Dining > Bakeware > Bread Pans" — every segment is a candidate synonym. The deepest node is usually the most specific and the most useful in a bullet; the shallower nodes are broader topical signals for the title.
2. Refinement filters in the left rail of the SERP. Material, type, capacity, finish, intended-use — every filter label is a controlled-vocabulary term Amazon has decided shoppers use. These map directly to attributes you should fill in your flat-file and reference verbatim in bullets.
3. Top-seller titles for the category. Pull the first two pages of organic SERP for the category's main query. Read the titles. The patterns repeat — "stoneware loaf pan with handles", "non-stick loaf pan set of 2" — and the repeating substrings are the synonyms the category's best-performing listings have already validated.
4. "Customers also bought" and "Frequently bought together" rails. The product names in those rails are tightly related to your category by Amazon's collaborative filtering. Synonyms that recur across multiple rails are high-confidence candidates.

The five-minute walkthrough

1. Open the SERP for the most generic query for your product ("loaf pan" not "stoneware bread loaf baking pan").
2. Note the full breadcrumb on a top-3 organic listing.
3. Read every left-rail filter group label and the values within each.
4. Copy the first 20 organic titles into a working document.
5. Open the top-3 detail pages and scan the "Customers also bought" rail.

You'll typically end up with 40–80 raw candidate phrases in five minutes — before touching any external tool.

Cleaning the raw list

• Strip brand names. "Le Creuset loaf pan" → "loaf pan". Brand-prefixed competitor terms aren't legal to use on your listing.
• Strip qualifiers that don't apply. "non-stick loaf pan" only stays if your product is actually non-stick.
• Keep stem variants. "loaf pan", "loaf pans", "bread loaf pan", "loaf baking pan" — all stay until Episode 14 evaluation.
• Group by intent. Material words, use words, size words, audience words. Grouping early makes the evaluation pass much faster.

The category-as-source limitation

The category surface only gives you the synonyms Amazon already knows about — the obvious ones. It won't surface the long-tail queries shoppers actually type but no current listing has captured. Episodes 04–13 cover sources that reach further into shopper intent (the title teardown of competitors, Google data, Sponsored Products data, the Amazon Search Term Report) — the category source is the starting point, not the destination.

The Vendor advantage

Vendor Central accounts can access the Brand Analytics dashboard, which surfaces the top search terms per category with click share and conversion share data. If you have Brand Analytics access, treat it as a sixth extraction surface for this episode — the top 100 search terms per browse node is a curated synonym list with built-in volume data, which is exactly what the evaluation episode (14) needs.