Commercial teams often use historical data to measure performance, assess progress toward revenue goals, and learn from past experiences with customers or prospects. And while this data provides helpful insight into the past, it can also be an invaluable tool in preparing for the future.
Propensity models allow commercial teams to predict customers’ future actions by looking at their past behaviors and relating those behaviors to internal and external variables. These models inform data-driven sales strategies and improve sales efficiency by giving sales leaders insights that help them:
- Prioritize accounts with the highest likelihood to buy, churn, or expand
- Balance opportunities across territories
In this blog, we’ll outline the various types of propensity models, the purpose of a propensity-to-buy model, and common mistakes to avoid as your organization develops or refines its propensity-to-buy model.
What is a Propensity Model?
A propensity model is a statistical prediction of the likelihood that a prospective or current customer will take a specific action. These predictive behavior models typically use logistic regression or other advanced non-linear machine learning techniques to model the relationship between multiple independent variables and the likelihood of a customer performing a certain action.
Propensity models use a combination of firmographic, technographic, psychographic, and historical data. They can be used to predict nearly any behavior — from engagement on a specific channel to movement across stages of the customer lifecycle.
Propensity modeling offers a repeatable, data-driven way to predict customers’ future behaviors while also identifying the triggers behind those behaviors. Sales teams can use these insights to inform their targeting efforts and strategic plays, like providing relevant product recommendations to speed up decision-making or taking proactive steps to reconcile with clients at risk of churning.
There are several types of propensity models that can be used to predict behaviors among existing customers. Examples include models that measure customers’ likelihood to:
- Refer prospective customers to the organization (Propensity to Refer)
- Accept upsells or cross-sells (Propensity to Upsell / Propensity to Cross-Sell)
- Stop doing business with the organization (Propensity to Churn)
Along with forecasting the behaviors of current customers, these predictive models help sales teams identify which prospective customers are most likely to buy their products and services via the propensity-to-buy model.
The Propensity-to-Buy Model
A propensity-to-buy model predicts how likely a prospect is to buy your product or service based on characteristics like their location, industry, and size, along with historical performance data from your organization. These historical data points account for variables like your past sales performance in certain markets, the pre-sales interactions that most often lead to closed deals, and how you perform against your competitors.
Propensity-to-buy models are critical in sales territory design and account planning, as they help commercial teams:
- Optimize sales capacity and coverage
- Prioritize and target accounts with the highest likelihood to convert
- Provide reps with an equitable number of high-potential accounts
What Does the Propensity-to-Buy Modeling Process Entail?
Each propensity-to-buy model requires a robust data set and sound statistical analysis to yield accurate results. Along with deciding what to include in your model, you’ll need to train and test your model to ensure it’s accurate before you start using it as a source of strategic insight. Here’s a high-level overview of the steps in the propensity modeling process:
- Build your model. Start by using logic to select and test variables that are likely to be related to a customer's propensity to buy your product or service. Align these variables with a combination of firmographic and technographic data points from customers that you won or lost in the past, including geography, industry, size, competitor presence, etc.
- Train your model. Use logistic regression or a more advanced non-linear machine learning algorithm to determine which variables are historically correlated with winning or losing an account. The algorithm identifies the relative strength (and combination) of variables that are most predictive of winning a new customer so that it can be used to predict the likelihood to buy for an account you haven't yet tried to sell to.
- Test your model before deploying it to your commercial teams. Start by testing your model on an independent holdout sample of historical data to see how accurate it is at predicting whether or not an account will be won, based on its firmographic, technographic, and other features. If your model generates propensity scores that align with your actual performance on those historical accounts, it’s ready to deploy. If your model is misaligned, you’ll likely need to reweight or even choose new variables to include.
- Deploy your model. Apply your propensity-to-buy model to prospective accounts you haven’t yet tried to close in order to generate a propensity score for each. These propensity scores can be used to inform sales prioritization, territory design, and account planning. You should plan to update your model with the latest sales performance data every six to twelve months.
Streamline Propensity Modeling With Coro’s MoneyMap
The success of your organization’s propensity modeling efforts hinges on the data you use to develop them and your commercial team’s ability to access and act upon your results.
Our MoneyMap tool provides your organization with proprietary data and advanced analytics capabilities through a single solution, so you can turn data about your customers, prospects, and competitive landscape into predictive models that inform revenue-focused decisions.
We’ll help you develop a custom MoneyMap that measures propensity to buy by product, territory, and customer segment and provide your sales reps with actionable ways to follow up with high-priority accounts. Then, our expert customer success team will partner with you to refresh your insights and propensity model as your needs evolve or the market shifts.
MoneyMap gives your entire commercial team a single source of truth when it comes to account prioritization, sales capacity planning, and other critical sales activities, empowering frontline reps and bringing your organization closer to its growth goals.
MoneyMap enables sales teams to:
- Identify cross-sell, upsell, and next-sell opportunities
- Inform capacity planning
- Expand into new verticals or market segments
- Segment accounts by propensity to buy
- Prioritize opportunities with the largest upside
Enterprises across the globe use MoneyMap to expand into new markets and realize their revenue goals in the most efficient manner possible.
See how one technology company powered its commercial growth strategy with MoneyMap and used the insights within it to realize ~$600M in total potential upside.
Subscribe to email updates
Additional content around the benefits of subscribing to this blog feed.