Onboarding Friction in SaaS Platforms: An Exploratory Qualitative-Computational Analysis

Abstract

Software-as-a-Service (SaaS) platforms face a persistent challenge in converting newly registered users into engaged adopters, but behavioural metrics alone provide limited evidence about the mechanisms of first-use friction. This study evaluates a transparent analysis workflow on 50 short, synthetically generated SaaS-onboarding interview records from an open educational repository. It therefore constitutes a methodological and theory-generating demonstration, not a study of observed users. Two coders applied the same eight-theme operational codebook, grounded in Cognitive Load Theory, the Technology Acceptance Model, Expectation Disconfirmation Theory, Goal-Setting Theory, and Self- Efficacy Theory. The resulting candidate typology comprises Information Overload (T1), Navigation Opacity (T2), Onboarding Discontinuity (T3), Personalization Deficit (T4), Goal-Means Misalignment (T5), Self-Efficacy Impairment (T6), Expectation Disconfirmation (T7), and Time-to-Value Friction (T8). Codebook coverage was reached by the seventh ordered record, while the remaining records supported prevalence estimation and exploratory subgroup comparisons.Inter-rater reliability was substantial on average (mean Cohen’s κ = 0.753). Large but imprecisely estimated associations appeared for Onboarding Discontinuity in Beginner-labelled records (OR = 63.00, 95% CI [6.52, 608.95], p < .001) and Expectation Disconfirmation in Advanced-labelled records (OR = 9.67, 95% CI [2.14, 43.56], p = .004). Because the corpus is synthetic, these patterns are hypotheses for validation with authentic users rather than population estimates. The contribution is a reproducible, deliberately modest workflow that links interpretive coding, transparent rule-based text scoring, and uncertainty-aware exploratory analysis.