You have learned about overrides but want to learn best practices on the overrides that will make the most sense in your model? Follow the guidelines of this article to get examples and tips 🎲 

Prerequisites

• You have an Architect or Admin HG Insights role

What is an override?

Overrides are rules you can add to your model to restrict the scores of your leads and contacts based on demographics, firmographics, or technographics traits.

Overrides can:

• boost the score of leads

• downgrade the score of leads

• avoid the score of leads falling beyond or going beyond a threshold

Override downgrading the score of leads

The idea is to make sure certain categories of leads will always be scored low whatever their other defining traits. This is particularly relevant for leads that you don't want to send to your sales team. Here are some classic overrides our customers usually create:

• Score spam emails low

• Score student emails low

• Score leads from your own company low

• Score leads from your competitors low

• Score leads from 'low GDP per capita countries' low

Override boosting the score of leads

The idea is to make sure certain categories of leads will always be scored good or very good whatever their other defining traits. This is particularly relevant for categories of leads that you're newly targeting or some of the big companies in the market that may have not performed well in the past but that you still want your sales team to talk to. Here are some recommendations:

• Score fortune_1000 emails at least good or very good

• Score target account emails at least good or very good

• Score leads from 100,000-employee companies at least good

When boosting segments of leads, also try to think about your ICP or certain specific technologies that your product integrates with. You don't want to miss out on these!

Override boosting an ICP persona

To create such overrides, the best way is to first create a persona computation and then add an override focused on this computation. Here's how to create a persona computation.

Operational data vs. Fit overrides

Missing phone numbers or hard-bounced emails make a lead harder to contact, but they don’t make the company itself less likely to convert.

We therefore recommend filtering on these fields inside your MQL workflow (e.g. “Fit = Very Good and phone_present = true”) rather than forcing a downgrading override in the Customer-Fit model.
If you do need to incorporate the check inside the model:

1. Expose the field in your CRM and let HG Insights pull it.

2. Map it in Attribute Mapping and create a Boolean computation (e.g. has_phone, email_invalid).

3. Reload the model, add a downgrading override, and save.

4. Use Override Impact Analysis to verify the distribution before deployment. support.HG Insights.com
   This keeps the predictive model focused on conversion potential while your CRM/MAP workflow handles data-quality gating.