AI Product Descriptions: Workflow That Sells

December 25, 2024AI • E-commerce • Marketing

E-commerce product descriptions

Loading...

Stores pay for product descriptions that remove doubt and drive clicks. This workflow turns specs and reviews into clear, benefit‑led bullets and SEO‑ready copy, delivered as a tidy CSV the merchant can paste directly. Scope stays tight; quality stays high.

Table of contents

  1. Intake and research (10 minutes)
  2. Prompt system (bullets, titles, descriptions)
  3. SEO checklist and character counts
  4. Delivery: CSV schema the store can import
  5. QA checklist and compliance
  6. Pricing tiers and bundles
  7. Outreach script
  8. FAQs

Loading...

Intake and research (10 minutes)

Prompt system

Bullets (benefit‑first)
Turn SPECS [+ top review phrases] into 5 bullets ≤ 180 chars.
Rules: each bullet starts with a benefit; include 1 objection‑handling bullet; plain language.
Description (100–150 words)
Summarize benefits for TARGET CUSTOMER; tone = BRAND VOICE; include materials/fit where relevant.
Rules: avoid superlatives; no unverifiable claims; 1 gentle CTA.
Titles (3 variants)
Create 3 concise title options with primary keyword and key attribute (e.g., material/color).

What is a prompt in an LLM (and why it matters here)?

A prompt is the instruction and context you give a language model to perform a task. Good prompts reduce ambiguity so outputs are usable without rewrites. For product copy, that means clear roles, constraints, evidence, and an output format that pastes cleanly into your CMS or CSV.

Use case: lifting a low‑performing collection

A store’s “Everyday Tees” collection had thin, duplicate bullets and vague descriptions. Rewriting 20 SKUs with this workflow increased CTR by 11% and reduced returns on two SKUs after adding specific fit/material guidance. The key wasn’t magic words—it was prompts that forced evidence and enforced counts, delivered in a clean CSV the merchant imported in one go.

Loading...

Use case: SEO‑first rewrite to win intent

A seasonal "Waterproof Hiking Jackets" collection failed to surface for the head term and long‑tails ("breathable rain jacket men", "lightweight waterproof shell"). We applied an SEO‑first prompt (below) across 15 SKUs and the category page, keeping copy human‑readable while aligning with search intent and marketplace rules.

Before

After (6 weeks)

Ready‑to‑use prompt (SEO‑first, product set)

Role: Senior e‑commerce SEO editor for an outdoor apparel brand.
Goal: Produce SEO‑first product copy that can rank for primary and secondary intents while matching brand voice.
Inputs:
- SKUs: [paste structured specs: material, membrane, weight, fit, features]
- Reviews: [top phrases/objections]
- Intents: [primary head term], [2‑4 long‑tails]
- Prohibited claims: [list]
- Internal link targets: [/size-guide, /care, /jackets/waterproof]
Constraints:
- Title: <= 60 chars; include key attribute; no adjectives pile‑ups
- Bullets: 5 items; <= 180 chars each; benefit‑first; 1 objection‑handling bullet
- Description: 100–150 words; mention material/fit/care where relevant
- Meta: title <= 60, description <= 155; natural, non‑spammy
- Language: plain, non‑technical where possible; no unverifiable claims
Output (JSON for each SKU):
{
  sku, title, bullets: string[5], description, keywords: string[],
  metaTitle, metaDescription, h1,
  faqs: [{ q, a }],
  internalLinks: string[] // slugs from Inputs
}
Checks:
- E‑E‑A‑T: cite specific materials/fit; avoid generic superlatives
- Duplication: bullets unique across SKUs; flag collisions
- Compliance: reject prohibited claims; keep reading grade ≤ 8
- Usage: exact primary phrase ≤ 2 occurrences total

Paste the resulting JSON into your pipeline, or map the fields into a CSV using the schema below. Keep one “evidence” column internally with the review phrases you echoed—useful for audits and future refreshes.

SEO checklist and character counts

Delivery: CSV schema

sku,title,bullets,description,keywords
1234,"Organic Cotton Crew Tee","• Soft, breathable comfort for all‑day wear | • Tailored fit feels polished without squeezing | • Durable stitching stands up to weekly washes | • Lightweight fabric keeps you cool on commutes | • Easy‑care cotton for grab‑and‑go outfits","Made from breathable organic cotton, this crew tee balances softness with a clean, everyday fit. Lightweight fabric keeps you comfortable while reinforced seams add durability. Pair with denim or joggers. Wash cold, hang dry.","organic cotton t‑shirt, breathable tee, everyday crew"

Hand off as CSV/Sheet. Include a second tab with character counts so clients see compliance at a glance.

QA checklist and compliance

Loading...

Pricing tiers and bundles

Outreach script

Subject: Quick win for [Store] product pages
Noticed duplicate/short listings on [collection]. I can deliver 10 SEO‑ready descriptions (title + bullets + 100–150w desc + keywords) in 48h for $179 flat. Includes CSV and character counts. Want 3 SKUs free to review?

FAQs

Can you match brand voice? Yes—provide 2 listing examples; I mirror tone and phrasing patterns.

How do you ensure uniqueness? I write benefit‑first lines tied to real specs and review language per SKU.

What about compliance? I avoid restricted claims and follow marketplace guidelines for phrasing and counts.

Next steps

Loading...

Real‑world use case: Write 10 SKUs for an Etsy shop

Translate specs and reviews into benefit bullets.

  1. Mine reviews
  2. Draft bullets ≤200 chars
  3. QA compliance and size/fit

Expected outcome: CSV ready to import; higher CTR and fewer returns.

Implementation guide

  1. Paste specs and 5 top reviews; extract benefits.
  2. Write 5 bullets ≤200 chars; add one objection answer.
  3. Fill CSV columns (title, bullets, desc, keywords).

Prompt snippet

Turn specs [PASTE] + top reviews [PASTE] into 5 benefit bullets (≤200 chars). Address top objection.

SEO notes

Loading...