LLM-driven Constrained Copy Generation through Iterative Refinement

2025-04-14

AI Summary

Marketing copy generation faces challenges in simultaneously satisfying multiple stringent constraints—including length, keyword inclusion, word order, and tone—while incurring high manual labor costs and suffering from poor generalization in existing approaches. Method: This paper proposes the first end-to-end large language model (LLM)-based iterative refinement framework specifically designed for multi-constrained copy generation. It innovatively integrates constraint-driven closed-loop optimization, multi-stage prompt engineering, and an online multi-armed bandit evaluation mechanism to jointly model and dynamically correct diverse complex constraints within a single unified pipeline. Contribution/Results: Experiments demonstrate a 16.25–35.91% improvement in constraint satisfaction rate. In A/B testing, generated copy achieves a 38.5–45.21% higher click-through rate than human-written baselines, significantly enhancing both business performance and system scalability.

Technology Category

• Natural Language Processing: Generation
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition
• Search and Optimization: Learning to Search

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Crafting a marketing message (copy), or copywriting is a challenging generation task, as the copy must adhere to various constraints. Copy creation is inherently iterative for humans, starting with an initial draft followed by successive refinements. However, manual copy creation is time-consuming and expensive, resulting in only a few copies for each use case. This limitation restricts our ability to personalize content to customers. Contrary to the manual approach, LLMs can generate copies quickly, but the generated content does not consistently meet all the constraints on the first attempt (similar to humans). While recent studies have shown promise in improving constrained generation through iterative refinement, they have primarily addressed tasks with only a few simple constraints. Consequently, the effectiveness of iterative refinement for tasks such as copy generation, which involves many intricate constraints, remains unclear. To address this gap, we propose an LLM-based end-to-end framework for scalable copy generation using iterative refinement. To the best of our knowledge, this is the first study to address multiple challenging constraints simultaneously in copy generation. Examples of these constraints include length, topics, keywords, preferred lexical ordering, and tone of voice. We demonstrate the performance of our framework by creating copies for e-commerce banners for three different use cases of varying complexity. Our results show that iterative refinement increases the copy success rate by $16.25-35.91$% across use cases. Furthermore, the copies generated using our approach outperformed manually created content in multiple pilot studies using a multi-armed bandit framework. The winning copy improved the click-through rate by $38.5-45.21$%.

Problem

• LLM-driven copy generation struggles with multiple intricate constraints
• Manual copy creation is time-consuming and lacks personalization scalability
• Iterative refinement effectiveness for complex copy constraints remains unclear

Innovation

• LLM-driven iterative refinement for copy generation
• End-to-end framework for multiple complex constraints
• Improved success rate and click-through performance