# How can I discard values when detecting duplicates?

## How to Handle Frequent Default Values That Cause False Positives in Duplicate Detection

When a field commonly contains a **default or placeholder value** (like `[Not Provided]`), it can cause **false positives** in your duplicate detection process. These values trigger unnecessary scoring and generate noise in your Delpha results.

This article explains how to use the **Discard Values** feature in Delpha to prevent such false positives and reduce background processing.

## Problem Example: \[Not Provided]

Imagine you receive Leads from a webform where only First Name and Email are required. If fields like **Last Name** or **Company Name** are missing, they default to `[Not Provided]`.

Your duplicate detection rules rely on:\
✅ Name + Email + Company

Now consider:

* 1,000 Leads all have **Last Name = \[Not Provided]**
* Each of these matches the others based on this placeholder value
* Result: 1,000 x 1,000 = **1,000,000 scoring operations**
* Many will clear the screening phase and be treated as **potential duplicates** — even though they’re not!

This clogs your system with **false positives** and wastes processing capacity.

## Solution: Use Discard Values

Delpha allows you to **ignore specific field values** during the duplicate detection process using the **Discard Values** setting.

### Step-by-Step:

1. Open the **Delpha Setup** app
2. Go to the **Duplicate Settings** tab
3. Scroll to the **Properties** section
4. Locate the **Discard Values** input
5. Add placeholder values (e.g., `[Not Provided]`, `N/A`, `Unknown`) as a **comma-separated list**

{% hint style="warning" %}
**Important**: No quotes, no spaces — format as:\
`[Not Provided],N/A,Unknown`
{% endhint %}

### What It Does:

* These values are **ignored during scoring**
* They are treated as **empty**, which disqualifies them from matching logic
* This helps reduce **false positives** and keeps your scoring meaningful

### Example Outcome

| Before Discard Values                    | After Discard Values       |
| ---------------------------------------- | -------------------------- |
| 1,000,000 unnecessary scoring operations | ✅ Reduced                  |
| High false-positive rate                 | ✅ Clean, focused detection |