# How does the Duplicate scoring work

## How Is the Duplicate Score Calculated in Delpha?

Understanding how Delpha computes the **Total Weighted Score** helps you fine-tune duplicate detection accuracy and control what records are flagged as potential matches.

## Formula

The total score is a **weighted average** of all matched fields:

```
Total Weighted Score = (Score₁ * Weight₁ + Score₂ * Weight₂ + ... + Scoreₙ * Weightₙ) / Total Weight
```

This score determines the similarity percentage between two records.

### Score Impact Factors

#### **Mandatory Checkbox**

Each field can be marked as **Mandatory** for evaluation:

* **Not Mandatory** (default):\
  If either record has an empty value, that field is excluded from the score.
* **Mandatory**:\
  Field is always used, even if one or both records have missing data. This can **lower the overall score** due to the missing information.

#### Example Scenarios

**Scenario 1**:\
**Field**: `LinkedIn URL`\
**Mandatory**: ✅\
**Both Records**: Empty\
➡ Score is penalized due to missing data.

**Scenario 2**:\
**Field**: `LinkedIn URL`\
**Mandatory**: ✅\
**Both Records**: Not empty\
➡ Score calculated normally using fuzzy or exact logic.

{% hint style="warning" %}
Use the **mandatory setting** for high-impact fields to strengthen matching accuracy, especially for strategic identifiers like email, LinkedIn URL, or phone.
{% endhint %}

###

<figure><img src="/files/rLNey1HUR3zZKiwY7Nit" alt=""><figcaption><p>Dulicate Scoring Fields</p></figcaption></figure>

### **Example**&#x20;

<table data-header-hidden><thead><tr><th width="135"></th><th></th><th></th><th width="135"></th><th></th></tr></thead><tbody><tr><td>Record</td><td><p>Name </p><p>[Weight = 100]</p></td><td><p>Email</p><p>[Weight = 100]</p></td><td><p>Account.Name</p><p>[Weight = 200]</p></td><td><p>LinkedIn URL</p><p>[Weight = 100]</p></td></tr><tr><td>A</td><td>Samantha</td><td><a href="mailto:samantha@cogicog.com">samantha@cogicog.com</a></td><td>Cogicog</td><td> </td></tr><tr><td>B</td><td>Sama<mark style="color:orange;">m</mark>tha</td><td>-</td><td>Cogicog</td><td> </td></tr><tr><td>Matching %</td><td>90%</td><td>0%</td><td>100%</td><td>0%</td></tr><tr><td>Partial Score</td><td><mark style="color:green;"><strong>Optional</strong></mark><strong>: 0.9*100</strong></td><td><p><mark style="color:red;"><strong>Mandatory</strong></mark><strong>: 0*100</strong></p><p><mark style="color:green;"><strong>Optional</strong></mark> </p><p><strong>Ignored</strong></p></td><td><mark style="color:green;"><strong>Optional:</strong></mark><strong> 1*200</strong></td><td><p><mark style="color:red;"><strong>Mandatory</strong></mark><strong>: 0*100</strong></p><p><mark style="color:green;"><strong>Optional</strong></mark><strong>: Ignored</strong></p></td></tr></tbody></table>

* With Email & LinkedIn = Optional \
  -> (0.9\**100* + 1\*200) / (100 + 200) = 96.7%
* With Email & LinkedIn = Mandatory \
  -> (0.9\**100 + <mark style="color:red;">0\*</mark>*<mark style="color:red;">100</mark> + 1\*200 + *<mark style="color:red;">0\*</mark>*<mark style="color:red;">100</mark>) / (100 + <mark style="color:red;">100</mark> + 200 + <mark style="color:red;">100</mark>) = 58%

{% hint style="info" %}
[More on the models used by Delpha for Duplicate scoring](/delpha-for-salesforce/how-to-faq/delpha-duplicate/what-algorithms-are-used-by-delpha.md)
{% endhint %}

&#x20;


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://help.delpha.io/delpha-for-salesforce/how-to-faq/delpha-duplicate/how-does-the-duplicate-scoring-work.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.