# What Kind of Language Does Humanizer's "Sales Language" Pattern (Pattern 4) Target?

> Humanizer's Sales Language pattern (Pattern 4) targets marketing phrasing buzzwords and hyperbolic adjectives removing promotional text to create neutral prose.

- Repository: [Siqi Chen/humanizer](https://github.com/blader/humanizer)
- Tags: deep-dive
- Published: 2026-09-07

---

**The "Sales language" pattern detects and removes marketing-style phrasing, buzzwords, and hyperbolic adjectives that make text read like promotional material rather than neutral prose.**

Humanizer's **"Sales language"** pattern—officially listed as **Pattern 16** in [`SKILL.md`](https://github.com/blader/humanizer/blob/main/SKILL.md)—is specifically engineered to strip away advertisement-style wording and replace it with concise, fact-focused descriptions. According to the [blader/humanizer](https://github.com/blader/humanizer) source code, this pattern targets three core characteristics of promotional writing that undermine neutrality in technical documentation and straight prose.

## What the Sales Language Pattern Detects

### Buzzwords and Hyperbolic Adjectives

The pattern flags **inflated descriptive language** that adds style without substance. In [`SKILL.md`](https://github.com/blader/humanizer/blob/main/SKILL.md) lines 42-45, the watch list includes terms like:

- "vibrant"
- "groundbreaking"
- "breathtaking"
- "must-visit"
- "renowned"

These adjectives signal marketing intent rather than factual reporting. The pattern treats them as indicators that a rewrite is needed.

### Overly Grandiose Claims

The pattern catches **unsourced superlatives** that exaggerate importance without evidence. Lines 44-46 of [`SKILL.md`](https://github.com/blader/humanizer/blob/main/SKILL.md) identify claims that inflate the significance of places, products, or organizations beyond what concrete details support. These assertions make text sound like a brochure rather than a reliable source.

### Appeal-First Phraseology

Perhaps most distinctively, the pattern targets **constructions that emphasize emotional appeal over fact**. Examples from [`SKILL.md`](https://github.com/blader/humanizer/blob/main/SKILL.md) lines 46-49 include:

- "nestled in the heart of"
- "rich cultural heritage"
- "stunning natural beauty"

These phrases prioritize how wonderful something sounds rather than what it actually is.

## Before-and-After Examples

The following table demonstrates how Humanizer's sales language detection transforms promotional text into neutral prose:

| Situation | Before (sales language) | After (Humanizer-cleaned) |
|-----------|------------------------|---------------------------|
| **Travel guide** | "Nestled within the breathtaking region of Gonder in Ethiopia, Alamata Raya Kobo stands as a vibrant town with a rich cultural heritage and stunning natural beauty." | "Alamata Raya Kobo is a town in the Gonder region of Ethiopia." |
| **Product description** | "Our groundbreaking app offers a seamless, robust experience that revolutionizes how you manage your workflow." | "The app helps you manage your workflow." |
| **Company promo** | "Experience our world-renowned, cutting-edge solutions that deliver unparalleled value and foster lasting partnerships." | "Our solutions help deliver value and build partnerships." |

In each case, the pattern removes promotional adjectives and reconstructs the sentence around core factual claims.

## Implementation Details

The sales language pattern is defined across three key files in the repository:

| File | Purpose |
|------|---------|
| [`SKILL.md`](https://github.com/blader/humanizer/blob/main/SKILL.md) | Contains the complete pattern definition, including the watch list and problem description at lines 42-49 |
| [`README.md`](https://github.com/blader/humanizer/blob/main/README.md) | Documents overall Humanizer usage and pattern philosophy |
| [`scripts/validate-package.py`](https://github.com/blader/humanizer/blob/main/scripts/validate-package.py) | Enforces metadata consistency and pattern numbering integrity |

The validation script ensures that Pattern 16 remains correctly indexed and that its detection rules stay synchronized with the documented behavior.

## When to Apply This Pattern

Use the sales language pattern when editing:

- Technical documentation that should remain vendor-neutral
- Encyclopedia entries and factual references
- Academic or journalistic writing
- Product descriptions that have drifted into marketing territory

The goal is **neutral readability**: text that tells readers *what* something is, not *how wonderful* it sounds.

## Summary

- The **"Sales language"** pattern (Pattern 16) targets **marketing-style phrasing, buzzwords, and hyperbolic claims** as defined in [`SKILL.md`](https://github.com/blader/humanizer/blob/main/SKILL.md).
- It detects **three language types**: inflated adjectives, unsupported superlatives, and appeal-over-fact constructions.
- The pattern **replaces promotional veneer** with concise, factual statements.
- Examples show reductions of 50-75% in word count while preserving core meaning.
- Implementation spans pattern definition in [`SKILL.md`](https://github.com/blader/humanizer/blob/main/SKILL.md), documentation in [`README.md`](https://github.com/blader/humanizer/blob/main/README.md), and validation via [`scripts/validate-package.py`](https://github.com/blader/humanizer/blob/main/scripts/validate-package.py).

## Frequently Asked Questions

### What specific words does Humanizer's sales language pattern watch for?

The pattern monitors buzzwords including "vibrant," "groundbreaking," "breathtaking," "must-visit," and "renowned" according to [`SKILL.md`](https://github.com/blader/humanizer/blob/main/SKILL.md) lines 42-45. These terms signal marketing intent that the pattern flags for removal.

### Is the sales language pattern actually Pattern 4 or Pattern 16?

While referenced as Pattern 4 in some contexts, the skill file lists it as **Pattern 16**. The [`scripts/validate-package.py`](https://github.com/blader/humanizer/blob/main/scripts/validate-package.py) validation script maintains correct pattern numbering to prevent drift between references.

### What makes sales language different from other Humanizer patterns?

Unlike patterns targeting jargon, redundancy, or complexity, the sales language pattern specifically addresses **persuasive intent**. It removes wording designed to influence reader opinion rather than wording that merely obstructs clarity.

### Can I customize which words the sales language pattern detects?

The current implementation uses a fixed watch list defined in [`SKILL.md`](https://github.com/blader/humanizer/blob/main/SKILL.md). To modify detection terms, you would edit this file directly and re-run the validation script to ensure pattern integrity remains intact.