How Many Keywords Should a Stock Photo Have? (Adobe Stock Best Practices)

Ask ten contributors how many keywords a stock photo needs and you'll get ten answers. The platforms don't help — Adobe Stock allows up to 49 keywords, Shutterstock up to 50, and neither explains how the count changes your visibility. Here's what actually matters.

The Short Answer

35–45 keywords for Adobe Stock, with the first 10 chosen like they're the only ones that exist. Below ~25 keywords you're leaving discoverability on the table. Above ~49 you're just padding — and padding can actively hurt you.

Why Order Beats Count

Adobe Stock's search engine weights your keywords by position. The first keyword carries the most ranking power, the tenth carries far less, and keyword #40 is close to decorative. This has a practical consequence most contributors miss:

Your first 10 keywords should be the exact phrases a buyer would type — not clever variations of them.

If your photo is a golden retriever puppy sleeping in a sunlit room, the first keywords should be golden retriever, puppy, dog, sleeping, cute — not adorable, domestic animal, canine companion, fur baby.

The Keyword Pyramid

A well-keyworded asset has three layers:

1. Core subject (1–8): the literal thing in the frame — objects, people, actions. These match buyer searches 1:1.
2. Context & attributes (9–25): setting, mood, colors, style, composition — sunset, beach, tropical, vacation, horizontal.
3. Concepts & use-cases (26–45): what the image means — freedom, new beginning, summer travel, copy space, banner background. These long-tail terms have less competition and quietly accumulate downloads over years.

Skipping layer three is the most common mistake. Conceptual keywords are how an ordinary beach photo keeps selling for a decade.

When More Becomes Worse

Keyword stuffing is detectable — and it backfires:

• Irrelevant keywords violate platform guidelines and can get assets rejected or accounts flagged.
• Duplicate-variation padding (dog, dogs, doggy, doggie) dilutes your ranking signal instead of adding to it.
• One-word spam in single-keyword fields (Shutterstock's arrangement) reads as noise to the matching engine.

If a keyword wouldn't help a buyer find or select your image, it doesn't belong on the list.

Exact Counts by Platform

| Platform | Recommended | Hard limit | Notes | | --- | --- | --- | --- | | Adobe Stock | 35–45 | 49 | Order matters; first 10 carry the most weight | | Shutterstock | 25–50 | 50 | Keywords typed individually; order matters less | | Freepik | 15–30 | 30 | Fewer, more precise terms win | | Vecteezy | 15–25 | 15+ | Subject-first |

Doing This at Scale

Hand-writing 40 ordered keywords per file is where most contributors give up — and where quality collapses. An AI metadata generator that understands the hierarchy (subject → context → concept) and orders keywords accordingly changes the math completely. Generate the draft automatically, then spend your 30 seconds reviewing the top 10 — that's the 80/20 of stock metadata in 2026.

The contributors who treat keywords as a ranking system — not a chore — are the ones whose year-old uploads still sell every week.