Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Recontextualizing Famous Quotes for Brand Slogan Generation

Ziao Yang, Zizhang Chen, Lei Zhang, Hongfu Liu

TL;DR

The paper tackles advertising slogan fatigue by introducing a paradigm that recontextualizes persona-related famous quotes for slogan generation. It presents a modular remix framework comprising quote matching, structure decomposition, vocabulary replacement, and remix generation, coupled with a short-text post-training step to bias the model toward slogan-length outputs. Evaluations across automatic diversity/novelty metrics and human judgments show improvements in variety, creativity, and emotional impact over strong LLM baselines, demonstrating better alignment with brand personas. The approach emphasizes interpretability and controllability, offering a scalable alternative to end-to-end slogan generation that leverages the depth of famous quotes while mitigating repetition and generic phrasing in real-world advertising systems.

Abstract

Slogans are concise and memorable catchphrases that play a crucial role in advertising by conveying brand identity and shaping public perception. However, advertising fatigue reduces the effectiveness of repeated slogans, creating a growing demand for novel, creative, and insightful slogan generation. While recent work leverages large language models (LLMs) for this task, existing approaches often produce stylistically redundant outputs that lack a clear brand persona and appear overtly machine-generated. We argue that effective slogans should balance novelty with familiarity and propose a new paradigm that recontextualizes persona-related famous quotes for slogan generation. Well-known quotes naturally align with slogan-length text, employ rich rhetorical devices, and offer depth and insight, making them a powerful resource for creative generation. Technically, we introduce a modular framework that decomposes slogan generation into interpretable subtasks, including quote matching, structural decomposition, vocabulary replacement, and remix generation. Extensive automatic and human evaluations demonstrate marginal improvements in diversity, novelty, emotional impact, and human preference over three state-of-the-art LLM baselines.

Recontextualizing Famous Quotes for Brand Slogan Generation

TL;DR

The paper tackles advertising slogan fatigue by introducing a paradigm that recontextualizes persona-related famous quotes for slogan generation. It presents a modular remix framework comprising quote matching, structure decomposition, vocabulary replacement, and remix generation, coupled with a short-text post-training step to bias the model toward slogan-length outputs. Evaluations across automatic diversity/novelty metrics and human judgments show improvements in variety, creativity, and emotional impact over strong LLM baselines, demonstrating better alignment with brand personas. The approach emphasizes interpretability and controllability, offering a scalable alternative to end-to-end slogan generation that leverages the depth of famous quotes while mitigating repetition and generic phrasing in real-world advertising systems.

Abstract

Slogans are concise and memorable catchphrases that play a crucial role in advertising by conveying brand identity and shaping public perception. However, advertising fatigue reduces the effectiveness of repeated slogans, creating a growing demand for novel, creative, and insightful slogan generation. While recent work leverages large language models (LLMs) for this task, existing approaches often produce stylistically redundant outputs that lack a clear brand persona and appear overtly machine-generated. We argue that effective slogans should balance novelty with familiarity and propose a new paradigm that recontextualizes persona-related famous quotes for slogan generation. Well-known quotes naturally align with slogan-length text, employ rich rhetorical devices, and offer depth and insight, making them a powerful resource for creative generation. Technically, we introduce a modular framework that decomposes slogan generation into interpretable subtasks, including quote matching, structural decomposition, vocabulary replacement, and remix generation. Extensive automatic and human evaluations demonstrate marginal improvements in diversity, novelty, emotional impact, and human preference over three state-of-the-art LLM baselines.
Paper Structure (9 sections, 3 equations, 3 figures, 6 tables)

This paper contains 9 sections, 3 equations, 3 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Method
  4. Experiments
  5. Experimental Setup
  6. Examples of Slogan Remix
  7. Performance by Quantitative Metrics
  8. Human Evaluation
  9. Conclusion

Figures (3)

  • Figure 1: Framework of recontextualizing famous quotes for brand slogan generation.
  • Figure 2: Slogan diversity metrics across eight domains. Each column corresponds to a diversity metric. The top row reports results for the appliance, baby, beauty, and electronics domains, while the bottom row reports results for the nutrition, household, furniture, and clothing domains.
  • Figure 3: Comparison of first-glance hook strength of our method against baselines. Radar plots visualize the relative hook strength of slogans generated by our method compared with three strong baseline models: GPT-4o, DS-L, and DS-Q. Each axis corresponds to one product domain, and values represent the normalized win ratio of our framework over the baseline, aggregated across all brand--persona combinations within that domain.