> For the complete documentation index, see [llms.txt](https://docs.sellermagnet.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.sellermagnet.com/selling-tools/listing-optimizer-overview/listing-quality.md).
# Listing Quality
{% hint style="info" %}
**Difficulty:** ๐ก Intermediate ยท **Reading time:** \~12 min
{% endhint %}
{% hint style="warning" %}
**Availability:** Listing Quality is part of the **Listing Manager AI** suite, which rolls out **per account**. While yours is pending you'll see an amber hourglass next to it in the **Listing-Manager** menu and a "you're on the list" note when you open it - at no charge until you decide to keep it. It needs a connected Amazon account. See [Listing Optimizer](/selling-tools/listing-optimizer-overview.md).
{% endhint %}
## ๐ Overview
**Listing Quality** is the main workspace of the Listing Manager AI suite. It scores every product in your catalog, tells you exactly which parts of each listing are weak and why, lets you fix them by hand or with AI, and pushes the approved changes back to your live Amazon listing through the Amazon API - no copy-pasting into Seller Central.
Where the classic [Listing Optimizer](/selling-tools/listing-optimizer-overview.md) analyzer audits one ASIN at a time and leaves the editing to you, Listing Quality is built to work across your whole catalog and to close the loop: analyze, edit, push, verify.
Every listing carries an **LQI Score (Listing Quality Index) from 0 to 100**. The LQI is simply the **share of checks the listing passes** - the backend counts how many checks passed out of the total and expresses it as a percentage. If a listing passes 30 of 37 checks, its LQI is roughly 81. Because the score is a plain pass rate, each check you fix moves the number by a predictable amount, and the failing checks tell you precisely where the points are.
### Tiers
Each LQI Score maps to a tier, shown as a colored badge in the table and detail panel:
| LQI Score | Tier | Color |
| --------- | ------------- | -------- |
| 80-100 | **Excellent** | ๐ข Green |
| 60-79 | **Good** | ๐ข Green |
| 40-59 | **Average** | ๐ Amber |
| 0-39 | **Poor** | ๐ด Red |
### Severity labels
Each individual check is labeled by how much it matters, so you can triage the failing ones highest-impact-first:
| Severity | Meaning |
| ------------ | ---------------------------------------------------------------------------------------------------- |
| **Critical** | A policy or truncation problem that can suppress the listing or hide your keywords - fix these first |
| **Moderate** | Hurts readability, ranking, or conversion but will not get you suppressed |
| **Minor** | Cosmetic or style polish |
| **Pass** | The check succeeds - nothing to do |
### Sub-scores
Alongside the overall LQI, two sub-scores split the listing into its two most-edited halves so you can see at a glance which one is dragging you down:
* **Title Compliance Score** - how well the title follows Amazon's title rules (the 25 title checks below).
* **Bullet Quality Score** - how well the bullet points are written and formatted (the 12 bullet checks below).
***
## โ
The 37 Quality Checks
The LQI is built from **37 checks**: **25 on the title** and **12 on the bullet points**. On top of those, Listing Quality also reads several **catalog signals** - images, description, A+ Content, video, review count, and rating - which appear as their own columns and contribute to the picture even though the core check count is the title-plus-bullet set.
Each failing check in the detail panel shows a plain-language reason (**"Why this matters"**) and how many quality points you recover by fixing it.
### Title checks (25)
{% tabs %}
{% tab title="Length & truncation" %}
| Check | What it looks for |
| ---------------------- | ---------------------------------------------------------------------------------------------------------------------------------------- |
| **Character Length** | Title in the 80-150 range; over-long titles get truncated or suppressed |
| **Desktop Truncation** | Rendered pixel width fits desktop search - the title starts getting cut off on desktop at around 1100px, with a warning from about 900px |
| **Mobile Truncation** | Rendered pixel width fits mobile search, where most Amazon traffic is - the title is cut off past about 580px |
| **Keyword Density** | Enough descriptive keywords (top titles use 6-10) |
| {% endtab %} | |
{% tab title="Policy & compliance" %}
| Check | What it looks for |
| ------------------------ | ----------------------------------------------------------------------- |
| **All-Caps** | No all-caps words (a common suppression trigger) |
| **Promotional Language** | No "sale", "best seller", "free shipping" and similar prohibited claims |
| **Amazon Trademark** | No other brand names that risk an IP complaint |
| **Regulatory Phrases** | No restricted or claim-like wording |
| **URLs / Email / Phone** | No contact details, which are a policy violation in a title |
| **Price / Currency** | No price - it changes and Amazon prohibits it |
| **Internal Codes** | No ASIN, SKU or UPC codes leaking into the title |
| {% endtab %} | |
{% tab title="Characters & formatting" %}
| Check | What it looks for |
| ------------------------ | ---------------------------------------------------- |
| **Forbidden Characters** | None of the characters Amazon rejects in titles |
| **Emojis & Symbols** | No emoji or symbols (stripped or rejected by Amazon) |
| **HTML Tags / Entities** | No stray HTML code |
| **Number Formatting** | Numerals rather than spelled-out numbers |
| **Punctuation** | No ellipses, semicolons or unmatched brackets |
| **Extra Whitespace** | No double spaces |
| **Leading / Trailing** | No leading/trailing whitespace or punctuation |
| {% endtab %} | |
{% tab title="Style & structure" %}
| Check | What it looks for |
| -------------------------- | ----------------------------------------------------- |
| **Title Case** | Consistent, professional title case |
| **Brand & Structure** | Leads with the brand and separates attributes clearly |
| **Consecutive Duplicates** | No repeated adjacent words |
| **Keyword Repetition** | No keyword repeated (Amazon indexes it once) |
| **Condition Terms** | No condition words that belong in the condition field |
| **Placeholder Text** | No leftover placeholder text |
| **Typo Detection** | No obvious misspellings |
| {% endtab %} | |
| {% endtabs %} | |
### Bullet checks (12)
| Check | What it looks for |
| ----------------------------------- | ---------------------------------------------------------------- |
| **At least 5 bullet points** | Listings with 5 bullets convert better |
| **Each bullet โฅ 150 chars** | Very short bullets waste the highest-attention block on the page |
| **Each bullet โค 250 chars** | Over-long bullets collapse behind "read more" |
| **Total chars โค 1000** | Past \~1000 characters Amazon stops indexing the rest |
| **Start with uppercase** | Bullets that start lowercase read as unfinished |
| **No emojis in bullets** | Amazon strips or rejects emoji |
| **No forbidden characters** | Certain characters can block the bullet update entirely |
| **Opens with a benefit** | Lead with the benefit, not the feature name |
| **States a measurable spec** | Give a concrete, measurable detail |
| **Answers fit or what is included** | Address compatibility or contents |
| **Names who it is for** | Say who the product is for |
| **Does not repeat itself** | No repetition across bullets |
{% hint style="info" %}
**Why the checks are worded around Amazon's own rules.** These are not generic writing tips - each check maps to a behavior that Amazon rewards or penalizes (indexing, truncation, suppression, IP). Clearing the Critical ones protects the listing from being hidden; clearing the rest lifts ranking and conversion.
{% endhint %}
***
## ๐ The Workspace
Open Listing Quality from the **Listing Manager** tab on the Listing Optimizer page. Scores are gathered automatically by a background worker, so the table is already populated when you arrive; you can also trigger a manual scan for any individual product.
### Stat cards
Four cards across the top summarize the catalog:
| Card | What it counts |
| ------------------- | -------------------------------------- |
| **Total Products** | Number of listings with quality data |
| **Avg. LQI Score** | Average LQI across those listings |
| **Needs Attention** | Listings below the "Good" threshold |
| **Critical** | Listings with critical-severity issues |
### The Product Listings table
The main table lists every scored product. Click any column header to sort, and click any row to open its detail panel.
| Column | What it shows |
| --------------- | -------------------------------------------------------- |
| **ASIN** | Amazon Standard Identification Number |
| **SKU** | Your Stock Keeping Unit |
| **Marketplace** | The Amazon marketplace the listing belongs to |
| **Title** | The current live title (with product thumbnail) |
| **LQI Score** | The 0-100 quality index, as a colored badge |
| **Bullets** | Bullet coverage (also the Bullet Quality Score on hover) |
| **Images** | Image count (HD count available in the detail) |
| **A+** | Whether the listing has A+ Content |
| **Video** | Whether the listing has a video |
| **Reviews** | Review count |
| **Rating** | Average star rating |
| **Last Scan** | When the listing was last scored |
| **Tier** | Excellent / Good / Average / Poor badge |
{% hint style="info" %}
Use the **Export CSV** option to download the visible table (respecting your current filters) for sharing with a copywriter or your team. Adjust the filters first - exporting with nothing in view returns nothing.
{% endhint %}
### Filters and search
Above the table:
* **Marketplace filter** - All Marketplaces, or a single marketplace.
* **Score filter** - All Scores, or one tier: Excellent (80-100), Good (60-79), Average (40-59), Poor (0-39).
* **Search box** - filter by ASIN, SKU, or title.
* **Scan Product** - pull fresh listing data from Amazon and re-score a product on demand.
***
## ๐ The Detail Panel
Click a product row to slide open its detail panel. A hero header shows the product image, title, ASIN, SKU, and current LQI Score, and the body is organized into four tabs.
{% tabs %}
{% tab title="Quality" %}
The **Quality** tab is the heart of the tool. It lists every check for this listing, sorted by severity, each with:
* a **pass or fail** status,
* a short **"Why this matters"** explanation,
* the **points** you would recover by fixing it,
* a rough **effort** estimate.
Below the checks sit the editable fields - **Title**, **Bullet Points**, **Description**, **Backend Search Terms**, **Images**, and **A+ Content** - each with a **Fix in editor** button and an **AI Optimize** button.
{% endtab %}
{% tab title="Keywords" %}
The **Keywords** tab tracks the search terms this listing targets and shows, per keyword, whether it appears in the **Title (T)**, **Bullets (B)**, and **Description (D)**. Add keywords manually or click **Extract** to auto-pull them from the listing content. Fields managed by A+ Content are marked as locked (not editable here).
{% endtab %}
{% tab title="History" %}
The **History** tab has two parts: a **Score Trend** chart of the LQI over time, and a **Change Log** listing every field you have changed, when, and from what. The Change Log is also the source for reverting a change (see the safety model below).
{% endtab %}
{% tab title="AI Studio" %}
The **AI Studio** tab is the AI content and image workspace for this listing. It offers one-click optimization across fields and the **AI Image Studio**, which can generate product photos from 14 presets (White BG, Lifestyle, Home, Outdoor, Infographic, Scale, Close-Up, Multi-Angle, Packaging, Before/After, Gradient, Luxury Dark, Seasonal, Flat Lay) or clean up a reference photo you upload. The studio also takes **background**, **mood**, and **camera-angle** inputs (for example a Macro angle), and enforces a safety rule on every image - no people or faces, and no text, badges, or claims burned into the image.
{% endtab %}
{% endtabs %}
### Translate a listing
Beyond the per-field "Translate" mode, the detail panel has a dedicated **Translate listing** button that opens a modal to localize a whole listing at once. Choose a **Target language** (11 languages, including Dutch, Japanese, and Turkish), pick a **Target marketplace** (which sets the target language and the byte limits), and tick the **Fields to translate** - Title, Bullet Points, and Description are all on by default. Click **Start AI translation** and each field translates with its own progress indicator. This is a distinct multi-field, multi-marketplace flow, more than the single-field Translate mode. As with every other action here, nothing is sent to Amazon until you push the result.
### The editors
Each field opens its own editor:
| Field | Notes |
| ------------------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Title** | Free-text; re-checked against the 25 title rules as you edit |
| **Bullet Points** | One bullet per line; the 12 bullet checks apply |
| **Description** | Free-text; locked and read-only when the listing is managed by A+ Content |
| **Backend Search Terms** | Hidden search keywords, capped at **250 bytes**. Not shown to buyers - no commas needed, do not repeat words already in the title or bullets, and skip competitor brands. A live byte counter warns you as you approach the limit |
| **Images** | Upload, paste a URL, or generate with the AI Image Studio, per slot |
| **A+ Content** | Fetch from Amazon, add or edit modules, or assign an existing template |
***
## ๐ ๏ธ How to Use It, Step by Step
{% stepper %}
{% step %}
### Scan a product
Pick a marketplace and search your catalog, or click **Scan Product** to pull fresh data from Amazon. The scan fetches the listing, checks the images, and computes the score, ending on a **"Scan complete - LQI Score: N"** confirmation. (Background scores refresh periodically on their own, so you only need to scan manually when you want the very latest.)
{% endstep %}
{% step %}
### Review the failing checks
Click the product row to open the detail panel and stay on the **Quality** tab. Work down the failing checks from Critical to Minor. Each one tells you **why it matters** and **how many points** it is costing you, so you always know what to fix first.
{% endstep %}
{% step %}
### Fix - by hand or with AI
For any weak field, either click **Fix in editor** to edit it yourself, or click **AI Optimize** / **Generate with AI** to have SellerMagnet rewrite it. AI output always lands in a **draft editor first** - it is a suggestion you review and edit, never an instant change. The AI follows Amazon's style rules by construction (Title Case, character limits, no promotional language, no emojis or forbidden characters), and the result is re-checked against the 37 checks.
{% endstep %}
{% step %}
### Push to Amazon
When you are happy with the draft, open **Update listing on Amazon**. Tick the elements you want to send - **Title**, **Bullet Points**, **Description**, **Keywords**, **Images** - and click **Save** (or **Save & Submit for Approval** for A+ Content). SellerMagnet submits the selected fields to Amazon.
{% endstep %}
{% step %}
### Verify
After the push, re-open the product (or re-scan) to confirm Amazon accepted it and the LQI climbed. The **History** tab records the submission and lets you track its status.
{% endstep %}
{% endstepper %}
### The "Update listing on Amazon" dialog
When you push, the **Update listing on Amazon** dialog asks you to **choose which elements you would like to update**. You tick only the fields you want to send - you never have to push everything at once. Depending on the field, the confirm button reads **Save to Amazon** or **Save & Submit for Approval** (A+ Content goes through Amazon's approval flow). The **Submit All Changes to Amazon** button in the Listing Builder pushes every pending field in one go.
***
## ๐ The Push Safety Model
Listing Quality can change your live listings, so it is built to make every step explicit and reversible. Here is exactly what happens - stated honestly:
1. **AI output is a draft first.** Anything the AI writes lands in the draft editor. It is not sent anywhere until you open the push dialog and choose to send it.
2. **Every button tells you its consequence.** Actions are badged with one of two labels so you always know where a click lands:
| Badge | Meaning |
| --------------------------------------------- | ---------------------------------------- |
| **Applies to the draft - not sent to Amazon** | Local only - safe to experiment |
| **Publishes to your live Amazon listing** | This one goes to Amazon when you confirm |
3. **Amazon queues the submission - it is not instant-live.** A successful push means Amazon **accepted the submission into its queue** ("Changes submitted to Amazon and queued for processing"). The listing updates only after Amazon finishes validating it, and Amazon can still **reject** a change during that validation. If it does, SellerMagnet surfaces Amazon's own error or warning wording rather than a generic message.
4. **Every submission is logged and trackable.** Each push writes a Change Log entry (with the real previous value), and hands back Amazon's submission ID so its status can be tracked.
5. **Changes can be reverted.** Because the Change Log stores the value each field held *before* the change, you can **Revert** a field to restore the live listing to its earlier content. A **drift guard** protects this: if the listing has changed on Amazon since your edit (by Amazon, another contributor, or an auto-published experiment), the revert refuses and shows you both versions so you can decide, rather than silently overwriting newer content.
6. **A text-policy pre-flight blocks known-bad content.** Before a submission is sent, a content check scans the text for wording Amazon is known to reject. If it finds a blocker, the push is stopped up front - saving you a round trip and an Amazon error code to decode.
{% hint style="warning" %}
**Draft versus live - the buttons tell you which.** Nothing reaches Amazon until you explicitly open the push dialog, tick the elements, and confirm. Treat a push like submitting an edit in Seller Central, not an instant edit.
{% endhint %}
***
## โ๏ธ Configuration
Most configuration lives inside the detail panel and the AI popovers.
### Per-field AI options
When you optimize a field with AI, you can steer the result:
| Option | What it does |
| ------------------------ | ---------------------------------------------------------------------------------------------------- |
| **Mode** | Create (from scratch), Optimize (improve existing), Competitor (outrank a rival title), or Translate |
| **Tone** | Professional, Premium-Luxury, Conversational, or Technical |
| **Target Audience** | Optional - who the copy should speak to |
| **Keyword Emphasis** | How hard the AI leans on your target keywords |
| **Length** | How long the generated copy should run |
| **Target Keywords** | Keywords the AI should work in |
| **Negative Keywords** | Words the AI must avoid |
| **Custom Instructions** | Free-text guidance (tone, angle, audience) |
| **Backend Search Terms** | Managed in its own editor, capped at 250 bytes |
| **Fields to optimize** | For one-click optimization, tick which fields the AI should rewrite |
| **Marketplace** | Sets the language and byte limits used for the listing |
### AI Settings
Open **AI Settings** to control the AI itself:
* **Text AI Model** and **Image AI Model** - the specific model to use.
* **Token Usage** - a meter showing your remaining monthly allowance.
{% hint style="info" %}
**AI is included - no API key needed.** The AI runs on SellerMagnet's own models; you do not bring or pay for your own OpenAI or Gemini key. Generations draw from a monthly **credit allowance** included with your plan (shared with the [Bulk Sourcer](/selling-tools/product-sourcer-overview/bulk-sourcer.md)) and reset every billing cycle. AI Settings shows how many credits you have left.
{% endhint %}
***
## โ ๏ธ Caveats and Good to Know
* **Per-account rollout.** The Listing Manager AI suite, including Listing Quality, rolls out per account. While yours is pending you'll see an amber hourglass next to it in the Listing-Manager menu and a "you're on the list" note when you open it - at no charge until you decide to keep it.
* **Requires a connected Amazon account.** Scanning, pushing, and reverting all go through the Amazon API (SP-API), so a live Amazon connection is required.
* **Scores refresh periodically.** A background worker keeps LQI scores up to date on its own; manual **Scan Product** is for when you want the very latest number immediately.
* **Amazon has the final say.** A push is a submission, not a guarantee - Amazon validates asynchronously and can reject.
* **A+ / description locking.** When a listing's description is managed by A+ Content, the description editor is read-only; edit the content through the A+ Content tab instead.
***
## โ
Best Practices
1. **Work highest-impact-first.** Sort the table by LQI ascending, open your worst-scoring high-revenue listings, and clear the **Critical** checks before anything cosmetic.
2. **Fix policy checks before conversion checks.** A suppressed listing earns nothing - clear all-caps, promotional language, trademarks, and forbidden characters first.
3. **Treat AI output as a strong first draft.** Always read it in the draft editor before you push it live.
4. **Push selectively.** Send one field, confirm Amazon accepted it and the LQI moved, then send the next - rather than pushing everything blind.
5. **Use the sub-scores to target effort.** A low Title Compliance Score but a healthy Bullet Quality Score tells you exactly where the work is.
6. **Re-scan after pushing.** Confirm the change landed and the score climbed before moving on.
7. **Keep the Change Log clean.** If a push looks wrong once live, revert rather than layering another edit on top.
***
## โ FAQ
How is the LQI Score calculated?
It is the share of quality checks the listing passes, expressed as a 0-100 percentage. Pass 30 of 37 checks and the LQI is about 81. Because it is a plain pass rate, each check you fix moves the score by a predictable amount.
What is the difference between LQI and the Title Compliance / Bullet Quality scores?
The LQI is the overall score across all checks. The Title Compliance Score covers only the 25 title checks, and the Bullet Quality Score covers only the 12 bullet checks - two sub-scores that tell you which half of the listing needs the work.
Do my changes go live on Amazon immediately?
No. When you push, Amazon accepts the change into its queue and applies it after validating - and it can still reject a change during that step. SellerMagnet logs every submission, lets you track its status, and lets you revert.
Can I undo a change I pushed?
Yes. The Change Log stores the value each field held before your edit, so **Revert** restores the live listing to its earlier content. If the listing changed on Amazon since your edit, a drift guard stops the revert and shows you both versions so you can decide.
Does the AI need my own OpenAI or Gemini key, and does it cost extra?
No key needed - the AI runs on SellerMagnet's own models and is included in your plan. Usage draws from a monthly credit allowance (shared with the Bulk Sourcer). Pick the text model and image model, and see your remaining credits on the Token Usage meter, in AI Settings.
Why is my description editor read-only?
Because that listing's description is managed by A+ Content. Edit it through the A+ Content tab rather than the plain description editor.
What is the 250-byte limit on backend search terms?
Amazon caps hidden backend search terms at 250 bytes. The editor shows a live byte counter and warns you as you approach it - Amazon will truncate or reject anything over the limit. Do not repeat words already in your title or bullets, skip commas, and avoid competitor brands.
Why did Amazon reject my push?
Amazon validates submissions asynchronously and can reject on content or attribute rules. When it does, SellerMagnet shows Amazon's own error or warning wording. A built-in text-policy pre-flight also blocks known-bad content before it is ever sent, to catch the common cases early.
***
## โก๏ธ What's Next?
{% content-ref url="/pages/6tasvaZvbyMV13hfI0BP" %}
[Listing Agent](/selling-tools/listing-optimizer-overview/listing-agent.md)
{% endcontent-ref %}
{% content-ref url="/pages/taNm8GCbHtpSjd6CGv43" %}
[Keyword Research](/selling-tools/listing-optimizer-overview/keyword-research.md)
{% endcontent-ref %}
{% content-ref url="/pages/VXlGkA4PfkAz77id3Txp" %}
[Listing Optimizer](/selling-tools/listing-optimizer-overview.md)
{% endcontent-ref %}