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

# Lead Scoring conversions

> Use machine learning to assign quality scores to leads based on their characteristics and likelihood to convert.

<Note>
  Lead Scoring is an enterprise feature that requires collaboration with the SegmentStream data team for implementation and optimization.
</Note>

Lead Scoring is an advanced conversion type that uses machine learning to assign quality scores to your leads based on their characteristics and likelihood to convert. This helps you identify high-value prospects and optimize your marketing spend toward leads that are most likely to generate revenue.

Lead Scoring analyzes your lead data to identify qualified leads and estimates their potential deal value. This allows you to:

* Focus on qualified leads that are most likely to convert
* Optimize marketing campaigns toward lead quality, not quantity
* Improve ROI by focusing budget on channels that generate valuable leads
* Export qualified leads to advertising platforms for better targeting

## Getting started

<Steps>
  <Step title="Assess your data">
    Review whether you have the required lead and sales data.
  </Step>

  <Step title="Choose integration method">
    Determine if you can use a standard CRM integration or need custom data setup.
  </Step>

  <Step title="Contact SegmentStream">
    Reach out to discuss your specific requirements and begin the setup process.
  </Step>
</Steps>

## Data requirements

There are two approaches depending on your CRM setup:

<Tabs>
  <Tab title="Standard CRM integration">
    If you use a well-known CRM system, SegmentStream can automatically connect to your existing data. For example, [HubSpot](/project-configuration/data-sources/hubspot) has a ready-to-use integration.

    Contact your SegmentStream manager to get details about connecting your specific CRM system. With standard CRM integration, SegmentStream automatically collects lead data and deal outcomes to train the scoring model.
  </Tab>

  <Tab title="Custom data integration">
    If you use a custom CRM or one not in the SegmentStream catalog, provide lead and sales data in the following format.

    **Leads data table**

    Required fields:

    | Field              | Description                                          |
    | ------------------ | ---------------------------------------------------- |
    | `created_at`       | When the lead was created                            |
    | `updated_at`       | When the lead was last updated                       |
    | `lead_id`          | Unique identifier for each lead                      |
    | `google_client_id` | Google Analytics Client ID when the lead was created |
    | `user_id`          | Internal user identifier that links to sales data    |

    Include as many relevant lead characteristics as possible to improve model accuracy: email, name, company name, company size, industry, job title, seniority level, lead source, country, annual revenue, budget, pain points, and product interest.

    **Sales/deals data table**

    Required fields:

    | Field        | Description                                 |
    | ------------ | ------------------------------------------- |
    | `created_at` | When the deal was closed                    |
    | `updated_at` | When the deal was last updated              |
    | `order_id`   | Unique identifier for each sale/deal        |
    | `value`      | Revenue amount of the deal                  |
    | `user_id`    | Links back to the lead data (same user\_id) |

    <Warning>
      To accurately predict customer lifetime value, include ALL purchases attributed to each `user_id` in the sales table, not only the first purchase. This helps the model understand the complete revenue potential of similar leads.
    </Warning>
  </Tab>
</Tabs>

## How Lead Scoring works

1. **Data analysis** -- SegmentStream analyzes your historical lead and sales data to identify patterns.
2. **Model training** -- Machine learning models learn which lead characteristics predict successful conversions.
3. **Qualification** -- New leads are evaluated and only those meeting quality thresholds become conversions.
4. **Value prediction** -- The system estimates potential deal value for each qualified lead.
5. **Daily updates** -- Lead qualification and scoring are updated daily as new data becomes available.

## What you get

* **Qualified lead conversions** -- only high-quality leads that meet configurable thresholds are created as conversions in SegmentStream.
* **Predicted deal values** -- estimated revenue potential for each qualified lead.
* **Exportable data** -- qualified lead conversions can be exported to ad platforms such as Google Ads and Meta for enhanced conversions, lookalike audiences, and bid optimization.
* **Reporting and analytics** -- use Lead Scoring conversions in SegmentStream reports to analyze qualified lead performance by traffic source and compare channel effectiveness.

## Best practices

* **Capture Google Client ID** -- ensure your forms capture the Google Analytics Client ID to link website behavior with lead data.
* **Collect rich attributes** -- the more relevant information you collect about leads at the time of creation, the more accurate the scoring will be.
* **Maintain data quality** -- keep lead and sales data clean and up-to-date.
* **Historical data** -- provide at least 6 months of historical data for optimal model training.

<Note>
  Attributes must be available immediately when the lead is created to ensure they can be used for next-day conversion processing. Attributes added later in your sales process may not be available for new lead scoring.
</Note>