> ## Documentation Index
> Fetch the complete documentation index at: https://docs.getsmelt.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Quality Scores

> Understanding file quality scores and data quality

Every uploaded file receives a quality score from 0-100, helping you understand your data's readiness for copy generation.

## How Quality Score Is Calculated

The score starts at 100 and deductions are made for:

| Issue                                     | Max Deduction   |
| ----------------------------------------- | --------------- |
| Empty rows                                | Up to 20 points |
| Incomplete columns (low avg completeness) | Varies          |
| Data quality issues                       | Up to 30 points |

## Factors That Increase Score

* More complete columns (fewer empty cells)
* Recognized fields present (name, email, company)
* Valid email formats
* No empty rows
* Critical fields have data

## Factors That Decrease Score

* Empty rows in your data
* Low column completeness (less than 50% filled)
* Invalid email formats
* Many empty cells
* Missing critical columns

## Score Ranges

| Score        | Rating    | Recommendation                          |
| ------------ | --------- | --------------------------------------- |
| **80-100**   | Excellent | Great data quality, proceed confidently |
| **60-79**    | Good      | Minor issues, should work well          |
| **40-59**    | Fair      | Review data, consider cleaning          |
| **Below 40** | Poor      | Clean data before processing            |

## Viewing Your Quality Score

After upload, the file detail page shows:

* **Overall quality score** (0-100)
* **Column-by-column analysis**
* **Detected field types**
* **Duplicate counts**

## How Quality Affects Results

<Note>
  Quality score reflects your **input data**, not Smelt's output quality.
</Note>

Higher quality input data leads to:

* Better personalization (more context for AI)
* Higher confidence scores on outputs
* Fewer generic results

Lower quality input data may result in:

* More generic outputs
* Lower confidence scores
* More flagged results

## Improving Your Quality Score

<Steps>
  <Step title="Remove Empty Rows">
    Delete any rows without meaningful data
  </Step>

  <Step title="Fill Critical Fields">
    Ensure name, company, and email columns are populated
  </Step>

  <Step title="Fix Invalid Emails">
    Correct obviously wrong email formats
  </Step>

  <Step title="Complete Partial Data">
    Fill in missing job titles, industries, locations where possible
  </Step>
</Steps>

<Tip>
  You don't need a perfect 100 score. Scores above 60 generally work well for most use cases.
</Tip>