Stock Image Metadata: How to Write Titles & Keywords That Sell

A great photo with weak metadata is invisible. Stock agencies and Google Images cannot "see" your picture the way a buyer does — they rely heavily on the title, keywords and description you give them. This guide covers how to write stock image keywords and titles that match what real buyers type, with a formula you can apply to every upload.

Think like the buyer, not the camera

The most common mistake is describing objects ("woman, laptop, table, window") instead of intent. Buyers search for what the image is for: "remote work", "online learning", "small business owner", "work from home India". Good metadata covers three layers:

1. Literal: what is visibly there — subject, action, setting, colours.
2. Conceptual: what it represents — success, teamwork, festival, sustainability.
3. Commercial: where it will be used — website banner, blog header, social ad, presentation background.

A title formula that works

Aim for a natural sentence of roughly 50–70 characters:

[Subject] + [Action] + [Location/Setting] + [Context]
Example: Young Indian woman working on laptop at home office during morning

• Lead with the main subject — it is what most people search first.
• Write for humans. A readable title converts better than a keyword list.
• Avoid brand names, trademarks and people's names unless you have the rights and releases.
• Do not repeat the same title across a series — vary the action or angle.

Keywords: order, count and quality

Most agencies accept up to about 50 keywords. More is not better — irrelevant keywords can hurt your ranking and annoy reviewers. A reliable structure:

• Top 10: the most important, most specific terms (subject, action, key concept). Some agencies give more weight to earlier keywords, so order matters.
• Next 15–20: setting, mood, colours, composition, demographics that are visually obvious.
• Last 10–20: broader concepts and use cases.

Use single words or short phrases as each agency recommends, include both singular and common plural forms where it matters, and never guess sensitive attributes such as religion or ethnicity unless they are clearly and respectfully part of the image.

Descriptions, alt text and slugs

If you also publish images on your own site, every image needs more than stock keywords:

• Meta description (up to ~155 characters) — a sentence that would make someone click the image page.
• Alt text — a short, literal description for screen readers and search engines. See our image SEO guide.
• Slug — lowercase, hyphenated, descriptive: woman-working-laptop-home-office.

Pre-upload checklist

• Title is a natural sentence with the main subject first.
• First 10 keywords are the most specific and relevant.
• No trademarks, logos or names without permission.
• No duplicate titles in the batch.
• Original filename preserved, so the CSV row matches the file exactly.
• Seasonal images tagged with the season or festival (Diwali, monsoon, wedding season) only when clearly visible.

Doing this for 100+ images

Writing all of this by hand takes 3–5 minutes per image. For a shoot of 200 photos, that is a full working day. The AI Image SEO Metadata Generator analyses each image with Google Gemini and fills 27 columns — title, categories, primary and secondary keywords, tags, trend and micro keywords, description, alt text, slug, mood and more — then exports one CSV. You still review the result, but you start from a solid draft instead of a blank cell. The workflow is explained in bulk image metadata with Gemini AI.