Keyword Research
Amazon Backend Keywords Limit: 249 Bytes Explained (2026)
The Amazon backend keywords limit is 249 bytes, not 250 characters. Here are the rules, byte math, and examples that actually index.
If you’ve been treating the Search Terms field like a 250-character text box, you’re probably losing indexing without knowing it. The real Amazon backend keywords limit is 249 bytes, not 250 characters, and the two aren’t the same once you add accented letters, emoji, or any non-ASCII symbol. Go over by a single byte and Amazon can drop indexing on the whole field.
This post walks through the exact byte math, what the field is for in 2026, and the specific rules that trip up most sellers.
What is the Amazon backend keywords limit in 2026?
The Search Terms field (what sellers call backend keywords or generic keywords) accepts up to 249 bytes of text per listing in most categories. Amazon’s current Seller Central Search Terms help page confirms the 249-byte cap and that going over causes the field to be ignored for indexing.
The confusing part is that Amazon’s seller UI often says “250 characters” in tooltips and old help pages, which is why so many sellers assume they have a 250-character text box. They don’t. The underlying field is measured in bytes, and the UI mislabels it in several places.
Some restricted categories (jewelry, certain pharma, some media types) have tighter caps or different field rules, so always check the category-specific listing guide in Seller Central before assuming 249 applies.
Amazon’s algorithms aren’t publicly documented in detail, so what follows reflects current public guidance and seller community observation, not insider access. Treat it as research, not legal advice, and if you’re in active enforcement for keyword abuse, talk to a qualified Amazon compliance specialist.
Why does Amazon measure bytes instead of characters?
Amazon runs on UTF-8 encoded text, where a “character” can be 1 to 4 bytes depending on what it is. A plain lowercase letter is 1 byte. An accented letter like é is 2 bytes. A Chinese or Japanese character is 3 bytes. An emoji is 4 bytes.
So “red shirt” is 9 bytes (8 letters plus 1 space). But “café rouge” is 11 bytes, not 10, because the é takes 2 bytes. Add a shrug emoji and you’re adding 4 bytes for a single visible glyph. This is why sellers writing in Spanish, German, or any language with diacritics hit the limit faster than English-only sellers.
If you’re curious how this maps to other listing fields, our writeup on Amazon character and byte limits across every field breaks down the full set, including titles, bullets, and A+ content constraints.
The byte-vs-character gap also matters when sellers paste from Word or Google Docs. Smart quotes, en-dashes, and non-breaking spaces are often multi-byte characters that look identical to their ASCII equivalents but silently eat your budget.
How do you actually count bytes in your Search Terms?
You can’t eyeball it. A field that looks like 240 characters in a text editor can be 280 bytes if it contains any special characters. The safest approach is to use a tool that counts UTF-8 bytes directly.
Our free Amazon character limit checker handles this for the Search Terms field specifically. Paste your draft keyword string, and it returns the exact byte count plus a warning if you’re over 249. Several Chrome extensions do the same thing, but if you’re already working in a listing tool, counting bytes natively is faster.
Here’s how common inputs stack up:
| Input | Characters | Bytes |
|---|---|---|
| red cotton shirt | 16 | 16 |
| café rouge chemise | 18 | 20 |
| 赤いシャツ | 5 | 15 |
| red shirt 👕 | 11 | 14 |
| smart quotes “red” | 18 | 22 |
The last row is the one that catches most sellers. Copy-pasting from Word turns straight quotes into curly ones (” ” instead of ” ”), and each curly quote is 3 bytes instead of 1.
What are the rules Amazon enforces on backend keywords?
Amazon’s Seller Central guidance lays out a short list of things that waste space or get the field ignored. These aren’t suggestions, they’re how the indexer actually processes the field.
- No repeated words. “Shirt red shirt cotton shirt” indexes the same as “shirt red cotton.” Repeating burns bytes without adding relevance.
- No punctuation. Commas, semicolons, periods, and dashes don’t help. Use single spaces between terms.
- No stop words. Articles and prepositions like “a,” “the,” “for,” “with” are stripped by the tokenizer. Don’t include them.
- No competitor brand names. Amazon’s prohibited search terms policy treats this as a listing violation that can lead to suppression.
- No subjective claims. Terms like “best,” “cheapest,” “new,” or “sale” are either filtered or risk policy flags under restricted content rules.
- No ASINs or product IDs. Amazon strips these.
- No profanity, abusive terms, or temporary statements. Season-specific claims like “Christmas 2024” age poorly and may violate current content guidelines.
If you want a deeper look at what’s banned versus just discouraged, our guide on Amazon prohibited keywords and claim risk covers the enforcement patterns we see most often.
Lower-case everything. The tokenizer is case-insensitive, so uppercase just wastes a human-visible signal (there isn’t one) without changing indexing.
What should go in the Search Terms field versus the title?
The field is most valuable for terms you can’t fit elsewhere in the listing. Everything already present in your title, bullets, description, or A+ content is indexed. Repeating those words in the backend adds zero ranking signal and burns byte budget.
Reserve the Search Terms field for:
- Synonyms your audience uses but you didn’t include in the title (“sofa” vs “couch”, “sneakers” vs “trainers”).
- Common misspellings of brand or product terms.
- Secondary-language terms for bilingual markets (Spanish terms on a US listing targeting Hispanic shoppers).
- Long-tail variants and modifiers that didn’t fit the character budget on the title or bullets.
- Use-case terms (“for travel”, “for small kitchen”, “for pet owners”) when those weren’t naturally covered elsewhere.
For how this fits into the full listing structure, our Amazon listing optimization guide walks through where each keyword type belongs. If you’re specifically rewriting titles, the product title optimization guide handles front-end placement, and the bullet points guide covers the middle layer.
Think of backend keywords as insurance against missed search terms, not a dumping ground for everything you want to rank for.
What are examples of good and bad backend keywords?
Here’s a concrete comparison for a hypothetical listing: a stainless steel insulated water bottle, 32oz, teal color.
Bad backend keywords (238 bytes):
best water bottle, insulated water bottle, stainless steel water bottle,
32oz water bottle, teal water bottle, hydro flask style, yeti style,
new 2024 water bottle, cheap water bottle, sale water bottle
What’s wrong: commas waste bytes, “water bottle” is repeated 7 times, “best,” “cheap,” and “new” are banned subjective terms, “hydro flask” and “yeti” are competitor brands that risk suppression.
Good backend keywords (229 bytes):
flask tumbler thermos canteen jug reusable eco friendly bpa free
leakproof vacuum double wall gym office travel hiking camping gift mom
dad teacher botella termo acero inoxidable
What works: no repeats, no commas, synonyms (flask, tumbler, thermos, canteen, jug), use cases (gym, office, travel, hiking, camping), gift intent (mom, dad, teacher), Spanish terms for bilingual search coverage, no competitor brands, no subjective claims.
The second version covers roughly 25 unique concepts in the same byte budget. That’s the goal: maximum unique tokens, zero waste.
How do you find keywords worth putting in the backend?
Backend keywords are leftover opportunity. You find them by first identifying every term your listing could rank for, then subtracting the ones already in your title, bullets, and description.
The practical workflow:
- Pull a full keyword list from a reverse ASIN lookup on 3 to 5 top competitors. This gives you the search terms Amazon associates with similar products.
- Cross-reference with Amazon Brand Analytics Top Search Terms if you have Brand Registry. This surfaces real search volume data from Amazon’s own system.
- Pull long-tail variants that are too specific or too long for the title but still have search intent behind them.
- Remove any term already in your title, bullets, description, or A+ content.
- Prioritize what’s left by estimated search volume, rank the remaining terms, and pack the top candidates into 249 bytes.
Our keyword research methodology post walks through this exact workflow in more detail, including how to use coverage indicators to spot keywords your listing is missing. The coverage indicators guide explains what those signals mean in a Keywords.am audit.
If you want a less manual approach, Keywords.am surfaces backend-eligible terms directly in the listing audit flow, flagging synonyms and long-tails that aren’t present elsewhere in your copy.
Frequently Asked Questions About Amazon Backend Keywords
What is the Amazon backend keywords character limit?
The limit is 249 bytes, which maps to about 249 characters if you stick to plain ASCII. Any accented letter, non-Latin script, or emoji counts as multiple bytes each, so your effective character budget shrinks. Going over the limit causes Amazon to ignore the field entirely for indexing.
Do backend search terms still help ranking on Amazon in 2026?
They still help, but their weight has dropped relative to title, bullets, and A+ content. Amazon’s catalog documentation confirms the field is indexed. The practical value is in capturing synonyms, misspellings, and secondary-language searches that don’t fit elsewhere, not in repeating primary keywords you’ve already placed up front.
Should I use commas or spaces between backend keywords?
Spaces. Amazon’s Seller Central guidance explicitly says commas and other punctuation waste byte budget without improving relevance. The field is tokenized on whitespace, so word order and separators between words don’t change what gets indexed.
What happens if I exceed the Amazon backend keywords limit?
Amazon typically stops indexing the entire field once it crosses the 249-byte cap. You lose relevance signal from every term in there, not just the overflow portion. Some sellers also report upload or save errors until the field is trimmed back under the limit.
Can I repeat keywords from my title in the backend?
You can, but there’s no indexing benefit. Each unique token is indexed once per listing regardless of how many fields it appears in. Reserve backend space for terms that aren’t anywhere else in your copy, since that’s where it actually earns its keep.
Conclusion
- The Amazon backend keywords limit is 249 bytes, not 250 characters. Count bytes, not characters.
- Spaces between terms, no commas, no stop words, no competitor brands, no subjective claims.
- Reserve the field for synonyms, misspellings, alt-language terms, and long-tails that don’t fit elsewhere.
- Repeating title or bullet words in the backend adds zero ranking signal.
- Audit your field against competitor keyword gaps every quarter, since search behavior shifts and new synonyms emerge.
If you’re guessing at your byte count or copy-pasting from Word and hoping for the best, that’s where listings lose indexing silently. Keywords.am audits every listing field for byte compliance, keyword overlap, and missed-synonym coverage, then shows exactly what to put in the Search Terms field based on what competitors rank for and you don’t.
Start your free Keywords.am trial and audit your backend keywords in under 5 minutes.