Revenue Potential Modeling for a B2B Software Provider
Estimating account-level upsell and cross-sell potential through customer peer comparisons, product-level models, and refreshable scoring.
Context
A B2B software provider needed to understand expansion potential within its existing customer base. Working through a growth strategy consultancy, we led the data science workstream within the broader consulting team. The objective was to estimate opportunity account by account and product by product, using differences in customer size, operations, and current purchasing.
Approach
We combined CRM records, transaction-level revenue, product information, and customer operating data into a modeling dataset. We tested variable transformations and normalized spending measures so customer comparisons reflected differences in scale. Nonparametric statistical tests, effect-size measures, and quantile regression helped assess which characteristics were useful for explaining spending patterns.
We grouped customers with comparable firmographic and operational characteristics, then used upper-performing peers to establish spending benchmarks. Product-specific models estimated headroom from deeper use of existing products and adoption of additional products. We documented the assumptions behind these estimates and built a reproducible R pipeline to refresh account scores as the underlying data changed.
Results
The consulting team received a bottom-up revenue potential model, product-level upsell and cross-sell estimates, and a documented scoring pipeline for the client’s revenue operations team. The outputs provided a consistent basis for comparing account opportunities and examining the assumptions behind them. These were modeled opportunity estimates, rather than realized revenue gains or a claim that every customer would reach the peer benchmark.
