Optimize Amazon backend search terms to maximize discoverability. By deduplicating, ranking, and formatting keywords not in the listing, generate an optimal 250-byte backend keyword set.
Supported platforms: Amazon、Shopify、WooCommerce、Walmart、TikTok Shop、Etsy、eBay、BigCommerce。
How to Use
优化我的后台关键词。我的产品是竹制笔记本支架。以下是我当前的标题和五点：[粘贴 listing]。以下是50个候选关键词：[粘贴关键词]。
Features
- 250-byte limit optimization (maximize keyword coverage with minimal space)
- Deduplication check against title, bullet points, and description
- Rank keywords by relevance and search volume signals
- Multilingual keyword inclusion strategy
- Misspelling and synonym coverage
- Prohibited term filtering (competitor brands, restricted claims)
- Before-and-after coverage comparison
How It Works
Step 1: Gather information from the user message — product, platform, current status, and goals.
Step 2: Ask all follow-up questions at once in multiple-choice format. Allow shorthand answers (e.g., "1b 2c 3a").
Step 3: Use the framework and methodology below for research and analysis.
250-Byte Limit Optimization System
- Byte calculation: English letters/numbers = 1 byte each. Chinese/Japanese = 2-3 bytes each. Special characters may also occupy multiple bytes. Use space-separated format (not commas/semicolons) to save separator bytes.
- Exact deduplication: Remove keywords already in the title or bullet points — they do not add extra indexing weight. Include only keywords not already visible in the listing content.
- Stem deduplication: The Amazon algorithm recognizes word roots. "Running" covers "runner" and "run". Merge variants into one optimal form to avoid redundancy.
- Semantic deduplication: Identify highly similar phrases. Keep the expression with higher search volume or broader coverage.
Keyword Ranking Algorithm
- Search volume weighting: Calculate a base score based on monthly search volume — higher search volume = higher priority.
- Competition correction factor: High-search-volume keywords usually have intense competition. Apply corrections for new products or products with few reviews. Boost medium-competition long-tail keywords with decent search volume.
- Relevance strength assessment: Only high-relevance keywords can bring valuable traffic. Remove low-relevance, high-search-volume terms (e.g., selling yoga mats but targeting "sneakers").
- Independent coverage value: Assess whether each keyword covers search demand that other keywords cannot cover. Low independent value = replaceable when space is tight.
Misspelling and Synonym Coverage
- Common misspellings: Include high-frequency errors (e.g., the common misspelling of "bluetooth," "blutooth"). Do not exhaustively include all variants — select only truly high-frequency errors.
- Synonym expansion: Cover regional variants and different consumer terms (e.g., in the US market, "portable charger" vs "power bank").
- Spanish keywords: Allocate approximately 30-50 bytes in the US market for Spanish translations of core selling points to reach Spanish-speaking consumers.
Prohibited Term Filtering
- Competitor brand filter: Never include other brands' trademarks — may cause the ASIN to be suppressed.
- Restricted claims filter: Prohibits "best", "number one", "free shipping" — violates Amazon policy.
- Subjective claims filter: Avoid "amazing", "greatest" — unverifiable claims.
- Punctuation cleanup: Remove extra spaces, repeated delimiters, special characters.
Data sources
- User provides the title, five bullets, and 50-100 candidate keywords.
- No external API. Deduplication and ranking are based on built-in logic.
Limitations
- 250 bytes is a hard limit — it cannot be exceeded.
- Duplicate keywords already present in the title/bullets waste bytes without adding indexing weight.
- Spelling variants consume bytes but can capture missed traffic.
Step 4: Provide structured, actionable output with specific recommendations, not vague guidance.
Output format
- Give a summary of findings first
- Include specific data points and benchmarks (if any)
- Provide prioritized action items
- Estimates based on incomplete data must be marked ⚠️
- End with specific next steps