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

# Quality Scoring

> How Data Legion's quality scores (high, moderate, low) help you assess the reliability of person data across phones, emails, locations, and social profiles.

How quality scores help you assess data reliability.

## Overview

Quality scores are available on person contact data (phones, emails, locations, socials). Each data point receives a confidence score that helps you understand data reliability and filter based on your use case.

<Note>
  Company data does not include per-field quality scores. Company records have a record-level `num_sources` count instead. See [Company Data Overview](/docs/company-data/overview) for details.
</Note>

## Quality Score Levels

| Score        | Description                                  |
| ------------ | -------------------------------------------- |
| **High**     | High quality data with strong verification   |
| **Moderate** | Moderate quality data with verification      |
| **Low**      | Lower quality data with limited verification |

## How Scores Are Assigned

We evaluate multiple factors to determine quality scores. The specific methodology is confidential, but scores reflect our assessment of data reliability.

## Using Quality Scores

### Filtering Strategies

**High-Stakes Operations** (Sales outreach, compliance):

* Filter for `confidence: high`
* Ensures maximum accuracy for critical use cases

**Full Coverage** (Research, analytics):

* Include `moderate` confidence data
* Balances coverage with quality

**Exploratory Analysis**:

* All confidence levels may be useful
* Use scores to weight results, not filter

### Example: Phone Quality

```json theme={null}
{
  "type": "mobile",
  "number": "+15551234567",
  "current": true,
  "last_seen": "2026-01-15",
  "num_sources": 3,
  "confidence": "high"
}
```

This phone has:

* High quality score
* Marked as current (recommended contact method)

This is the most reliable data you can get.

### Example: Email Quality

```json theme={null}
{
  "address": "jane.doe@company.com",
  "type": "professional",
  "current": true,
  "last_seen": "2026-01-20",
  "validated": true,
  "validation_status": "valid",
  "num_sources": 2,
  "confidence": "high",
  "hash_sha256": "a1b2c3d4e5f6...",
  "hash_sha1": "a1b2c3d4e5f6...",
  "hash_md5": "a1b2c3d4e5f6..."
}
```

This email has:

* Validated (deliverability confirmed)
* Marked as current (recommended contact method)
* High confidence score

## Common Questions

<AccordionGroup>
  <Accordion title="What does &#x22;confidence: high&#x22; actually mean?">
    It means we've assessed this data point as high quality based on multiple signals. It's a measure of our confidence in the data, not a guarantee of accuracy.
  </Accordion>

  <Accordion title="Why do some data points have moderate confidence even when they seem reliable?">
    We evaluate multiple factors to determine quality scores. Some data points may receive moderate scores based on various quality indicators.
  </Accordion>

  <Accordion title="How accurate is the scoring?">
    Our scoring is continuously refined based on data quality metrics. Scores reflect our best assessment of data reliability.
  </Accordion>
</AccordionGroup>

## Related Documentation

* **[Data Freshness](/docs/data-quality/data-freshness)** - Understanding data recency