— Amazon Ads Bulk Operations Guide

Creating a Sponsored Products Keyword campaign via bulk sheet

Build a Sponsored Products Keyword campaign where you specify every search term and match type. Full row-by-row walkthrough, match type strategy (exact, phrase, broad), bid patterns across match types, common mistakes, and a downloadable template you can fill in alongside.

The campaign type where you choose every search term

If Auto campaigns are about letting Amazon explore, Keyword campaigns are about exploiting what you already know. You specify the exact search terms you want to bid on, the match type for each (exact, phrase, or broad), and the bid for each. Amazon does no targeting decisions on your behalf — every impression goes through a keyword you explicitly added. That precision is what makes Keyword campaigns the workhorse of most mature Amazon Ads accounts, and the reason they're the natural follow-up to Auto campaigns once you have search-term data.

This article walks through building one from scratch via bulk sheet. If you haven't read the anatomy article yet, read it first — the column structure here assumes you know what Entity, Operation, and the ID columns do. The template generator on this page is pre-set to SP — Keyword campaign; download it alongside reading the walkthrough.

When to use Keyword campaigns

Three primary scenarios:

  • Harvesting converters from Auto campaigns. Your Auto campaign has been running for two to four weeks. The Customer Search Term report shows specific search terms that drove orders. Those terms are exactly what you put in a Keyword campaign — usually at exact match — so you can bid more aggressively on the converters without paying for everything Amazon's algorithm wants to test.
  • Branded campaigns. Bidding on your own brand name and product names. These campaigns typically have very high conversion rates (the searcher already knows you) and low CPCs (low competition for your exact brand). They protect against competitor conquesting and lock in branded traffic at favorable economics.
  • Competitor and category terms you researched. Specific keywords you identified through tools or domain knowledge — you know "12 quart air fryer" converts for your product because that's what your product is. You bid on it directly without waiting for Amazon to discover it.

Keyword campaigns are not great for discovery — you only learn about terms you put in. For new SKUs in unfamiliar categories, run an Auto campaign first, then harvest into Keyword once you have data.

The five rows you'll create

The minimal Keyword campaign for one SKU with two keywords needs five rows:

  1. Campaign row — the campaign itself.
  2. Ad Group row — one ad group inside the campaign.
  3. Product Ad row — links the SKU to the ad group.
  4. Keyword row — the first keyword.
  5. Keyword row — the second keyword. (Add more rows for more keywords.)

In practice, a real Keyword campaign rarely has just two keywords. A typical harvested campaign has 10–30 keywords, meaning 10–30 keyword rows below the campaign/ad group/product-ad rows. The structure scales: same three header rows, then one keyword row per keyword.

Row 1: the Campaign row

The campaign-level cells are almost identical to an Auto campaign, with one critical difference: Targeting Type is MANUAL, not AUTO. That single value tells Amazon this campaign uses explicit keyword (or product) targets rather than auto-discovery.

  • Campaign ID — unique value you make up. Used by every subsequent row. A descriptive pattern: kw-airfryer-exact-2026 or brand-defense-2026.
  • Campaign Name — display name in Seller Central. Often encodes the campaign type, match type strategy, and time period: SP-KW-AirFryer-Exact-2026.
  • Start DateYYYYMMDD format, no separators.
  • Stateenabled or paused.
  • Daily Budget — numeric, $1 minimum. Keyword campaigns often warrant higher budgets than Auto campaigns once you're past the harvest stage, because the traffic is more predictable.
  • Bidding StrategyDynamic bids - down only, Dynamic bids - up and down, or Fixed bid. Same options as Auto. For harvested keyword campaigns where you have conversion data, "Dynamic bids - up and down" is more often appropriate because Amazon has the conversion signal it needs.

Pre-filled: Product (Sponsored Products), Entity (Campaign), Operation (Create), Targeting Type (MANUAL).

Row 2: the Ad Group row

Ad groups in Keyword campaigns serve an important function: they're how you group keywords with similar bid strategies. Two patterns are common:

  • One ad group per match type. All exact-match keywords in one ad group, all phrase-match in another, all broad in a third. This lets you set different default bids per match type. Most common in harvest campaigns.
  • One ad group per theme. All keywords related to "air fryer capacity" in one ad group, all related to "air fryer features" in another. Better for very large keyword lists where you want budget-pacing differences by theme.

For most campaigns starting out, one ad group per match type is the right choice. Walkthrough assumes that pattern.

  • Campaign ID — match the Campaign row above.
  • Ad Group ID — unique value. Pattern: kw-airfryer-exact-2026-ag-exact.
  • Ad Group Name — display name.
  • Stateenabled.
  • Ad Group Default Bid — fallback bid if you don't specify a per-keyword bid. Numeric, $0.02 minimum.

Pre-filled: Product, Entity (Ad Group), Operation (Create).

Row 3: the Product Ad row

Identical to the Product Ad row in an Auto campaign. This is what links your SKU to the ad group so the campaign actually advertises a real product.

  • Campaign ID — match row 1.
  • Ad Group ID — match row 2.
  • Ad ID — unique, often the SKU plus -ad.
  • Stateenabled.
  • SKU — your exact seller SKU.

Rows 4 and beyond: the Keyword rows

This is where Keyword campaigns diverge from Auto. Each keyword you want to target gets its own row. The required cells per keyword row:

  • Campaign ID — match row 1.
  • Ad Group ID — match row 2.
  • Keyword ID — unique per row. Pattern: kw-airfryer-12-quart-exact.
  • Stateenabled or paused.
  • Bid — keyword-specific bid. Numeric, $0.02 minimum. Optional: if you leave this blank, Amazon uses the Ad Group Default Bid for this keyword. Many sellers leave it blank to keep bid management at the ad group level.
  • Keyword Text — the actual search term, e.g., 12 quart air fryer. Lowercase is standard; Amazon normalizes anyway.
  • Match Typeexact, phrase, or broad. Lowercase.

Pre-filled: Product, Entity (Keyword), Operation (Create).

Match types: what each one means and when to use it

Match type is the most important strategic choice in Keyword campaigns. The three types behave very differently:

  • Exact — your ad only shows for search queries that closely match your keyword, with allowances for plurals, common misspellings, and a few connector words. If your keyword is air fryer 12 quart, exact match will trigger on "air fryer 12 quart," "12 quart air fryer," "air fryers 12 quart," but not "12 quart air fryer for family." It's the most restrictive and the cleanest signal. Use it for harvested converters and any keyword where you want tight control over what triggers your ad.
  • Phrase — your ad shows for queries that contain your keyword phrase in the same order, with additional words before or after. Keyword air fryer phrase match triggers on "best air fryer," "air fryer with rotisserie," "air fryer for small kitchen," etc. Useful for broader exposure when you've identified a high-value root phrase but want to capture longer-tail variations.
  • Broad — your ad shows for queries Amazon judges to be related to your keyword, in any order, with any additional words, and including synonyms and related concepts. Keyword air fryer broad match might trigger on "convection oven," "kitchen countertop oven," or other related searches. Useful for discovery, but spending can scale unexpectedly. Use with conservative bids and active negative keyword management.

A common pattern is to add the same keyword in all three match types across three ad groups (or three campaigns), with different bids per match type. The same converter at exact match warrants a higher bid (precise intent, predictable performance) than broad match (variable intent, variable performance).

How to think about bids across match types

A typical starting bid structure for a harvested keyword:

  • Exact match: highest bid. You have conversion data, you trust this term. Bid to win the impression.
  • Phrase match: 60–80% of the exact match bid. Wider net, less predictable, lower bid.
  • Broad match: 30–50% of the exact match bid. Often used purely for discovery — finding new variations that you can then harvest into exact match.

For example, if your harvest data shows "12 quart air fryer" converts at $1.10 CPC for a 25% ACoS, you might bid $1.20 exact, $0.85 phrase, $0.50 broad. The exact match captures the proven converter aggressively; phrase and broad are exploratory.

Branded campaigns are different. Branded keywords typically have low CPCs (no competition for your brand name), so you can bid generously without exposure. Some sellers cap branded bids at $0.50 or lower because they win regardless.

Download a template and fill it in

Use the template generator on this page — it's pre-set to SP — Keyword campaign. The five-row structure (Campaign, Ad Group, Product Ad, and two example Keyword rows) shows what cells need filling. Duplicate the Keyword row pattern for each additional keyword.

Common mistakes specific to Keyword campaigns

  • Match Type capitalized. Amazon expects exact, phrase, broad — all lowercase. Exact with a capital E fails.
  • Mixing match types in one ad group at the same bid. Technically allowed, but you lose the ability to bid differently per match type. Most strategies benefit from separate ad groups per match type.
  • Forgetting that broad match scales unexpectedly. A broad keyword on a popular root term can drain a daily budget in hours if you set the bid too high. Start broad-match bids conservative and raise based on data.
  • Adding too many keywords at once. Uploading 200 keywords in one go means you can't tell which keywords are doing the work — you only see the aggregate. Better to start with a focused 10–20 and add more in subsequent uploads as you see performance.
  • Duplicate Keyword Text in the same ad group. If you upload two rows for "12 quart air fryer" both at exact match in the same ad group, Amazon rejects the second as a duplicate. Each (keyword text, match type) pair must be unique per ad group.
  • Stale harvests. Search term performance shifts over time. A converter from six months ago may not still convert. Refresh harvests every 30–60 days rather than uploading old data into a new campaign.

What happens after upload

Same as Auto campaigns: Amazon validates the sheet, the campaign appears in Seller Central within 5–30 minutes, and spending starts immediately if the state is enabled. Keyword campaigns typically reach steady-state faster than Auto campaigns because the targeting is fixed — Amazon doesn't need to "discover" anything, just bid your keywords when they're searched.

The data you'll want to look at after 7–14 days:

  • Per-keyword performance. Which keywords drove orders? Which spent money with no orders? The Keyword Performance Report (or downloading a fresh bulk sheet) shows this.
  • Per-match-type performance. If you separated by ad group, you can see whether exact is outperforming phrase, whether broad is converting at all, etc.
  • Search term breakdowns for broad and phrase keywords. Even though you targeted "air fryer" broad, the Customer Search Term report shows what actual queries triggered your ads. These are the next harvest candidates — high-converting actual queries you can promote to exact match.

Multi-keyword and multi-match-type patterns

A simple expansion: take one converting keyword and add it in all three match types. The bulk sheet has:

  • 1 Campaign row
  • 3 Ad Group rows (one for exact, one for phrase, one for broad — assuming the per-match-type strategy)
  • 3 Product Ad rows (one per ad group, referencing the same SKU)
  • 3 Keyword rows (same Keyword Text in each, different Match Type and Ad Group ID)

That's 10 rows for one keyword across three match types. For 20 keywords each across three match types, you're at 70 rows — still very manageable. This is the point where bulk sheets dramatically outperform clicking through the UI, where 70 rows would mean 70 separate forms.

Next steps

After your Keyword campaign has been running for two weeks:

  • Adjust bids based on per-keyword performance. Covered in the Bulk Bid Adjustments article. Bulk sheets are particularly powerful here because you can update dozens of keyword bids in one upload.
  • Add negative keywords for search terms that triggered your ads but didn't convert. Covered in the Negative Keywords article.
  • Promote new converters from broad/phrase match into exact match by harvesting again. Covered in the Search Term Harvesting article.

And the bigger picture: managing 5–10 SKUs each with Auto + Keyword + Product Target campaigns adds up to 15–30 campaigns. Each one with its own keywords, bids, negatives, and placement adjustments. BulkSheet Pro automates this entire structure — it takes a SKU list and produces all the campaigns at once, validates everything, and integrates the ongoing optimization workflow. Try it free →