From Words to Conversions: Leveraging Large Language Models for Dynamic Landing Page Generation

Abstract

In the realm of modern business, the demand for effective online marketing is undeniable. Nevertheless, the creation of landing pages for marketing campaigns poses a
fundamental question: Should the implementation be the domain of a web developer
or a marketing specialist? The interdisciplinary nature of this task, requiring a fusion
of coding and marketing skills, underscores a significant gap in current research. Addressing this gap, this thesis introduces a specialised AI landing page generator that
operates through natural language input, eliminating the need for coding expertise.
This empowers individuals without coding knowledge or marketing expertise to create
impactful landing pages independently.
To evaluate the system, usability tests were conducted followed by interviews with
marketing and web development professionals. The results demonstrated that business
specialists perceived the quality of AI-generated content positively. The AI provided a
basic structure and introduced new ideas, while human refinement is needed to ensure
the uniqueness of the content. In addition, the technical audits revealed high performance, accessibility and search engine optimisation of the landing pages generated.
Furthermore, comparisons between GPT-3.5 and GPT-4 revealed that GPT-4 generally
produced better quality content and was preferred by the participants for its creativity
and code quality. These findings highlight the potential of AI-driven tools to streamline the creation of landing pages, enhancing efficiency in digital marketing and web
development.

Date of Award	2024
Original language	English (American)
Supervisor	Andreas Stöckl (Supervisor)