Deal Signals: Building a Weighted Scoring Algorithm from Scratch
Designed and validated a deal-prediction scoring system with no existing playbook to follow
Manan Shah
Private Equity Intelligence Analyst at Gain.pro
From their time as
Private Equity Intelligence Analyst
Gain • 2025
Overview
Gain.pro was building an internal tool called Deal Signals, designed to track and predict upcoming investment deals. When Manan was brought onto the project, there was no established algorithm or methodology to follow. The question of how to weight signals and predict deal activity was open.
The Story
Gain.pro was building an internal tool called Deal Signals, designed to track and predict upcoming investment deals. When Manan was brought onto the project, there was no established algorithm or methodology to follow. The question of how to weight signals and predict deal activity was open.
He worked through the problem iteratively. He tested multiple approaches, evaluating each against deals that had already happened to see whether the model would have caught them. Several approaches failed that test: the score was not pointing toward deals that had actually closed, which told him the model was wrong.
He landed on a weighted scoring system that balanced multiple signals and produced results that aligned with historical deal activity. The approach was not obvious at the start; it emerged from testing and eliminating what did not work.
When he presented the approach to his manager, the manager was skeptical. Manan held his position, backed by the data he had built. The manager tested the scoring method independently and confirmed it was correct. The weighted scoring approach was adopted.