Product Description Generation Dataset Template

Overview

Product description generation datasets train AI models to create compelling, SEO-friendly product descriptions from structured product data — specifications, features, category, brand, and target audience. E-commerce companies with thousands or millions of SKUs need automated description generation that maintains brand voice consistency, highlights relevant features for each product category, and produces unique content that performs well in search engine rankings.

The training data maps structured product inputs (attributes, specifications, category metadata) to high-quality prose descriptions. Each training example shows the model what information to include, what tone to use, and how to structure the description for a specific product type. The dataset must cover your full product taxonomy — a description style that works for consumer electronics will not work for fashion apparel or home furnishings. Category-specific patterns in vocabulary, feature emphasis, and formatting must be well-represented.

Effective product description datasets also encode length and format preferences. Some categories need concise 50-word descriptions for mobile listings, while others benefit from detailed 200-word descriptions with bullet points for desktop pages. Including explicit format instructions in the training data teaches the model to adapt its output style based on the requested format, making a single model versatile enough for multiple content needs across your product catalog.

Dataset Schema

interface ProductDescriptionExample {
  instruction: string;
  input: string;        // Structured product attributes (JSON or formatted text)
  output: string;       // Generated product description
  metadata: {
    category: string;
    word_count: number;
    format: "short" | "standard" | "detailed" | "bullets";
    brand_voice: string;
  };
}

Sample Data

[
  {
    "instruction": "Write a compelling product description for an e-commerce listing. Use a professional yet approachable tone. Include key features and benefits.",
    "input": "Product: Wireless Noise-Cancelling Headphones\nBrand: SoundPeak\nModel: SP-NC400\nPrice: $179.99\nCategory: Electronics > Audio > Headphones\nFeatures: Active noise cancellation, 40mm drivers, Bluetooth 5.3, 35-hour battery, USB-C charging, foldable design, built-in microphone, multipoint connection (2 devices), weight 250g\nColors: Midnight Black, Arctic White, Navy Blue\nTarget: Commuters, remote workers, music enthusiasts",
    "output": "Escape the noise and immerse yourself in pure sound with the SoundPeak SP-NC400 Wireless Headphones. Advanced active noise cancellation blocks out ambient distractions — from busy commutes to open-plan offices — so you can focus on what matters.\n\nPowered by precision-tuned 40mm drivers, the SP-NC400 delivers rich, detailed audio across the full frequency range. Bluetooth 5.3 ensures stable, low-latency connectivity, and multipoint pairing lets you seamlessly switch between your laptop and phone without disconnecting.\n\nDesigned for all-day comfort and convenience, these headphones weigh just 250g and fold flat for easy storage. The 35-hour battery means you can go days between charges, and when you do need to recharge, USB-C fast charging gets you back to listening quickly. The built-in microphone delivers crystal-clear calls for your remote meetings.\n\nAvailable in Midnight Black, Arctic White, and Navy Blue.",
    "metadata": {
      "category": "electronics-audio",
      "word_count": 148,
      "format": "standard",
      "brand_voice": "professional-approachable"
    }
  },
  {
    "instruction": "Write a short, punchy product description for a mobile e-commerce listing. Maximum 60 words.",
    "input": "Product: Organic Cotton T-Shirt\nBrand: EcoWear\nMaterial: 100% GOTS-certified organic cotton, 180gsm\nFit: Relaxed fit\nFeatures: Pre-shrunk, double-stitched hems, tagless comfort label\nSizes: XS-3XL\nColors: 12 colors available\nTarget: Eco-conscious consumers",
    "output": "Made from 100% GOTS-certified organic cotton, the EcoWear tee combines sustainability with everyday comfort. Relaxed fit with pre-shrunk fabric means it fits perfectly from the first wear. Double-stitched hems for lasting durability. Tagless label for itch-free comfort. Available in 12 colors, sizes XS-3XL.",
    "metadata": {
      "category": "apparel-basics",
      "word_count": 48,
      "format": "short",
      "brand_voice": "eco-conscious"
    }
  }
]

Data Collection Guide

Start by collecting your best-performing existing product descriptions — those with high conversion rates, low bounce rates, and strong organic search rankings. Pair each description with the structured product data from your Product Information Management (PIM) system or product database. This creates training examples that teach the model your brand's proven description patterns.

For categories where you lack high-quality descriptions, commission your best copywriters to create template descriptions for 50-100 products per category. Provide them with the structured product attributes as input and ask them to write descriptions in your brand voice. These human-authored examples will define the style and quality standard for each category.

Ensure consistent formatting of the input data across all examples. Standardize how product attributes are presented — use consistent field names, formats, and ordering. The model learns input patterns from the training data, so inconsistent input formatting leads to unpredictable output quality. Ertas Data Suite can help standardize input formats during the data preparation pipeline.

Quality Criteria

Every product description in the training data must be factually accurate — specifications, features, and claims must match the product attributes in the input. Inaccurate training examples teach the model to hallucinate product features. Review each example to verify that no claims are made that are not supported by the input attributes.

Descriptions should be unique and avoid boilerplate language. If too many training examples share identical phrases or sentence structures, the model will produce generic, template-sounding output. Ensure variation in sentence structure, vocabulary, and feature ordering across examples within the same category.

Validate that descriptions meet your SEO requirements — they should naturally incorporate relevant keywords without keyword stuffing, use appropriate heading structures when applicable, and meet minimum and maximum word count targets for each format type. Test generated descriptions against your existing search rankings to verify SEO effectiveness.

Using This Template with Ertas

Import your product catalog data and existing descriptions into Ertas Data Suite. Use the data transformation tools to standardize input formats and pair product attributes with their descriptions. Apply PII redaction if product data contains supplier contact information or internal pricing data that should not appear in the training set. Export the prepared dataset in Alpaca format for fine-tuning.

After training in Ertas Studio, the GGUF-exported model can generate descriptions locally, which is particularly valuable for product catalogs containing pre-launch product details that must remain confidential until announcement dates.