Over the last 7 years of my career in Private Equity I’ve worked with over 60 companies—only 1 had already built a propensity model to prioritize accounts. In the dynamic world of marketing, business development & sales, not all accounts are created equal; propensity modeling is a powerful statistics-driven technique that transforms how companies prioritize and engage with potential customers.
All private equity funds are doing this, whether they call it “CapDb” (Vista), “MoneyMap” (Bain), or simply “propensity modeling” (TCV), the core principle remains the same: using data-driven insights to identify and prioritize accounts with the highest potential for purchasing or expansion.
Using sophisticated data science techniques and machine learning, we can predict the likelihood of a specific business outcome. In sales and marketing, this typically means:
💰 Identifying accounts most likely to buy
📈 Predicting potential for account expansion
🌍Optimizing territories
🎧Focusing GTM efforts on the most promising opportunities
When I analyzed closed-won accounts 12 months on from implementation, over 80% of new business had come from A accounts.
1. Data Collection and Integration
Successful propensity modeling relies on comprehensive data collection, starting with historical win & churn data from your company’s CRM.
That data is then enriched against an array of databases to provide firmographic and technographic insights beyond what most companies look at. For example, a potential correlative variable for an “A” account could be the percentage of developers in an account relative to peers.
On top of this, using AI has turbocharged our ability to see the impact of market trends on an accounts propensity to buy. An example here could be Trump’s tariff policies, we can calculate which industries and locations will face headwinds and tailwinds.
The inverse here is true too, churn data can be used to identify which accounts are less likely to stay customers for years to come, even should they become a client today.
2. Advanced Analytics Techniques
Modern propensity models utilize machine learning algorithms to identify hundreds of potentially significant variables, and evaluate them either individually or tandem to show multivariable correlation. Example: if an account has installed both Monday.com and Microsoft 365 does our software have a higher win rate?
3. Visualization and Actionability
The true power of propensity modeling lies in its ability to translate complex data into actionable insights. This often includes:
🗺️ Geographic heat maps - which locations should we be focussing on?
⏱️ Account priority dashboards - how do we prioritize the data we have?
⚖️ Rolled-up sales territory optimization visualizations
