In my research, I explored how conversational dynamics—specifically the depth and timing of topic discussions—impact conversion rates in alternative lending sales conversations. Using NLP techniques, including a fine-tuned BERTje model, I analyzed transcribed client interactions from Swishfund to uncover patterns that influence client decision-making.

I found that deeper discussions of certain product-centered topics, such as loan costs, were positively associated with conversion rates, while excessive focus on loan details tended to discourage clients. Timing analysis revealed that, contrary to my initial expectations, delaying discussions of financial specifics often aligned better with the natural flow of conversations, leading to higher conversion rates. Additionally, variables like the total number of transcripts and deal value emerged as significant predictors of conversion, emphasizing the importance of sustained client engagement and deal-specific factors.

This research highlights the potential of machine learning tools in uncovering nuanced conversational effects, offering actionable insights into tailoring communication strategies. By balancing information delivery and optimizing engagement, I aim to provide a framework for improving sales outcomes and advancing the understanding of interpersonal communication in both sales and psychological research contexts.