Fractional Attribution in Complex B2B Buying Journeys: Comparative Analysis of Multi-Touch Models and Pipeline Optimization for Enterprise Sales

Abstract

In complex B2B buying environments, enterprise sales require multi-touch, multi-channel journeys with multiple decision-makers, which make it difficult to know exactly which conversions can be attributed to which channels. Traditional models such as Last-Touch or Linear do not give proper credit to the customer journey, which can lead to an incorrect marketing and sales resource allocation. This study evaluates and compares several attribution models (Last-Touch, Linear, Time Decay, Shapley Value, Markov Chains and advanced probabilistic/algorithmic models) and proposes a powerful fractional attribution model accounting for touchpoint sequence, account-level influence and intensity of engagement. Multiple B2B enterprises' historical sales and engagement data were analyzed over 12-18 months and models were tested using indicators of Conversion Attribution Accuracy, Pipeline ROI Correlation, Engagement Balance, MAE, and RMSE. It shows that the state-of-the-art models (Shapley, Bayesian, Algorithmic/ML) reach higher accuracy (up to 84%) and performance metrics such as better ROI alignment (r = 0.78) and better engagement fairness (0.80) than the traditional ones. This is supported by an ablation study, which also demonstrates the critical importance of sequence and account-level influence on multi-touch journeys. The proposed framework for improved revenue attribution accuracy, optimized marketing budget allocation, and actionable points for enterprise sales strategies not only contributes to the research but also practical applications in B2B analytics.