---
title: "Add new enrichment sources to your fit segmentation"
slug: "add-new-enrichment-sources-to-your-fit-segmentation"
updated: 2025-08-07T12:12:31Z
published: 2025-08-07T12:12:31Z
---

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

# Add new enrichment sources to your fit segmentation

This guide provides a 3-week approach to build upon existing fit segmentation while adding custome enrichment sources. The 3-week timeline allows for thorough testing and validation while maintaining business continuity.

The methodology outlined here emphasizes validation at every step, cross-functional collaboration, and iterative improvement to deliver industry enrichment that drives measurable business impact.

## Prerequisites - Week 1

- Confirm the **source** of the new enrichment - usually Salesforce
- Confirm the **data points** to enrich
  - Company Number of employees
  - Company Industry
  - Company Country
  - Lead job title
- If **Industry** is included, create a spreadsheet **mapping table** to unify the different industry classifications from the different enrichment sources (MadKudu’s out-of-the-box industries vs your own Salesforce industries)
- Align on **fallback logic**
  - Define hierarchical priority, for example:
  - Salesforce **Account** Enrichment Provider **A** > Salesforce **Contact** Enrichment Provider **A** > Salesforce **Lead** Enrichment Provider **A** > Salesforce **Account** Enrichment Provider **B** > Salesforce **Contact** Enrichment Provider **B** > Salesforce **Lead** Enrichment Provider **B** > **MadKudu** out-of-the-box enrichment

## Step by step configuration - Week 2

- [**Pull the new enrichment**](/v1/docs/how-to-pull-additional-fields-to-use-in-the-madkudu-platform)**from all identified sources**
- [Create the **attribute mapping**](/v1/docs/attribute-mapping) for these new fields once they are pulled into MadKudu
- [Create **combined computations**](/v1/docs/combined-computations-leveraging-enrichment-from-madkudu-and-your-crm)****according to the decided priority
  - Name them “Employees combined”, “Industry combined”, “Country combined”
  - “Title combined” already exists by default; you just need to edit it by adding the new sources of enrichment
  - Find below all necessary templates, simply replace the attribute names

**employees_combined** - Type: Numeric - Mode: Advanced

```sql
CASE
	WHEN employees > 0 THEN CAST(employees AS NUMERIC)
	WHEN CAST(hg__employees_avg AS NUMERIC) > 0 THEN CAST(hg__employees_avg AS NUMERIC)
	WHEN CAST(salesforce_lead_numberofemployees AS NUMERIC) > 0 THEN CAST(salesforce_lead_numberofemployees AS NUMERIC)
	--WHEN CAST(salesforce_contact_employees AS NUMERIC) > 0 THEN CAST(salesforce_contact_employees AS NUMERIC)
	WHEN CAST(salesforce_account_numberofemployees AS NUMERIC) > 0 THEN CAST(salesforce_account_numberofemployees AS NUMERIC)
	ELSE NULL
END
```

**employees_combined_bucket** - Type: String - Mode: Advanced

```sql
CASE
	WHEN employees_combined = 0 THEN '1 - 0'
	WHEN employees_combined > 0 AND employees_combined <= 10 THEN '2 - ]0,10]'
  WHEN employees_combined > 10 AND employees_combined <= 100 THEN '3 - ]10,100]'
  WHEN employees_combined > 100 AND employees_combined <= 1000 THEN '4 - ]100,1000]'
  WHEN employees_combined > 1000 THEN '5 - ]1000,infinity['
END
```

**industry_raw_combined** - Type: String - Mode: Advanced

```sql
CASE
	WHEN industry IS NOT NULL THEN industry
	WHEN hg__top_level_industry IS NOT NULL THEN hg__top_level_industry -- it can be a number
	WHEN salesforce_lead_industry IS NOT NULL AND LOWER(salesforce_lead_industry) NOT LIKE '%unknown%' THEN salesforce_lead_industry
	--WHEN salesforce_contact_industry IS NOT NULL AND LOWER(salesforce_contact_industry) NOT LIKE '%unknown%' THEN salesforce_contact_industry
	WHEN salesforce_account_industry IS NOT NULL AND LOWER(salesforce_account_industry) NOT LIKE '%unknown%' THEN salesforce_account_industry
	ELSE NULL
END
```

**industry_combined**- Type: String - Mode: Advanced

Here you will replicate the mapping you prepared in the spreadsheet Week 1

```sql
CASE
WHEN LOWER(industry_raw_combined) LIKE '%automotive%' OR LOWER(industry_raw_combined) LIKE '%automobile%' THEN 'Automotive'
WHEN LOWER(industry_raw_combined) LIKE '%consumer%discretionary%' THEN 'Consumer Discretionary'
WHEN LOWER(industry_raw_combined) LIKE '%consumer%goods%' OR LOWER(industry_raw_combined) LIKE '%sporting%goods%' THEN 'Consumer Goods'
...
ELSE NULL
END
```

**country_combined**- Type: String - Mode: Advanced

```sql
CASE
	WHEN company__country IS NOT NULL THEN company__country
	WHEN pers_country <> 'not in table' AND pers_country <> 'cbit_unknown' THEN pers_country
	WHEN hg__hqcountry IS NOT NULL AND hg__hqcountry NOT IN ('115054','160004','196210','197110') THEN hg__hqcountry
	WHEN salesforce_lead_country IS NOT NULL THEN salesforce_lead_country
	WHEN salesforce_contact_mailingcountry IS NOT NULL THEN salesforce_contact_mailingcountry
	WHEN salesforce_account_billingcountry IS NOT NULL THEN salesforce_account_billingcountry
	ELSE NULL
END
```

- **Duplicate all live fit model(s)**
- **Replace**the original computations by combined computations in:
  - [Trees](/v1/docs/customer-fit-how-to-create-or-edit-a-decision-tree)
  - [Overrides](/v1/docs/customer-fit-overrides)
  - [Signals](/v1/docs/customer-fit-signals)
- Check the [performance](/v1/docs/evaluate-your-draft-customer-fit-performance-against-your-live-model) of the new model against the live model to make sure the **changes don’t have an impact** on segmentation volumes

## Iteration and Go Live - Week 2&3

- Model **validation**
  - Pair review the sample page several times to check that the new enrichment is showing up as expected
  - Iterate on the configuration if you spot any misconfiguration or any gaps
- **Deploy** the new model
  - Consider deploying on a Friday afternoon to avoid disrupting business
- **Monitor**the Data Quality Improvements:
  - You should witness an increase in data coverage
  - This means a reduction in "Unknown" signals
  - And a reduction in the mismatch between signals and CRM information

**Estimated time: 12 hours**