How do people make money with stock video?

Short answer: you upload video clips to stock marketplaces like Pond5 and Shutterstock, and earn a royalty — typically 30–40% of the sale price — every time a buyer licenses a clip. The money is slow at first, compounding, and heavily skewed: a small catalog earns almost nothing, and a large catalog of the right footage earns steadily for years.

Stock video is the quietest business on this list. There is no audience to build, no client to please, no algorithm to chase. You film something, upload it, tag it, and forget it. Then months later, someone making a commercial in another country buys the license, and a few dollars land in your account. Then it happens again. The whole model rests on one fact: buyers of video — advertisers, YouTubers, documentary producers — would rather pay $50–$200 for a ready-made clip than spend $5,000 shooting it themselves.

The royalty math

The numbers differ by platform, and they've moved against contributors over time, so get the current ones before you commit. On Pond5 — the largest marketplace dedicated to stock footage, with over 34 million clips — contributors currently earn 30% of the sale price for non-exclusive content and 40% for exclusive content (the rates were revised downward in early 2025). Pond5's distinctive feature is that you set your own prices, so a clip you price at $100 earns you $30–$40 per sale.

Shutterstock's contributor earnings run roughly in the 35–50% range depending on the licensing tier. Subscription platforms like Storyblocks and Artgrid pay differently — often a share of a subscription revenue pool rather than per-clip royalties, which means lower per-download payouts but steadier volume.

Put concretely: a clip that sells for $79 on Pond5 as a non-exclusive contributor puts about $24 in your pocket. That sounds small until you remember the clip took you an afternoon to shoot and it can sell dozens of times over a decade. One long-term contributor documented earning over 100,000 euros since 2006 from a catalog built steadily over the years. The per-sale math is modest. The compounding is the business.

What actually sells

The biggest beginner mistake is filming what looks beautiful instead of what buyers search for. Generic nature shots — sunsets, ocean waves, forest paths — are the most oversaturated category in every marketplace. They still sell occasionally, but you're competing with millions of similar clips.

What sells consistently, according to working contributors: business and workplace footage (meetings, offices, people working on laptops), lifestyle clips (families, cooking, fitness, everyday routines), healthcare and medical settings, technology close-ups, and diverse, authentic-looking people doing ordinary things. The buyers are mostly advertisers and content producers making things about modern life, and they need footage that looks like real modern life — not a postcard.

A useful mental model: shoot for the search bar. Buyers type phrases like "small business owner packing orders" or "nurse talking to patient" or "aerial view of suburban neighborhood at dusk." If you can picture the search query a buyer would type, and your footage is what they'd want to find, you're in the right territory.

The metadata nobody wants to do

Here's the unglamorous truth: a great clip with bad metadata earns nothing, because nobody finds it. Every marketplace runs on search, and search runs on the tags, titles, and descriptions you write. The people who earn well treat metadata as half the job.

The practical rules: write a descriptive title that matches how a buyer would describe the shot ("Timelapse of busy city street at night," not "city vibes 4k"). Add 10–50 keywords per clip, mixing general terms ("city," "night," "traffic") with specific ones ("timelapse," "downtown," "commute"). Tag in multiple languages if the platform allows it — buyers search in their own language. Choose the right categories. And be honest: misleading tags get clips rejected or buried, and buyers who feel tricked don't buy from you twice.

This is tedious. Everyone agrees. The contributors who succeed are the ones who do it anyway, clip after clip, because they understand that uploading without tagging is like opening a shop with no sign.

The honest timeline

Time to first payout typically runs 3 to 12 months. That is the single most important number in this article, because it's where most people quit. You upload 50 clips, nothing sells for four months, you conclude the business is dead, and you stop — right before the catalog would have started compounding.

The earnings arc follows a predictable pattern, documented by multiple contributors: a slow first year (one contributor reported net earnings of around 570 euros in under a year, including two individual Pond5 sales at $199 each), then monthly earnings climbing as the catalog grows (that same contributor reached about 500 euros a month by month thirteen). The curve bends upward only after the catalog reaches critical mass — generally a few hundred quality clips.

Treat weekly uploads as a non-negotiable habit rather than a one-time project. The contributors earning real money are the ones still uploading in month fourteen.

Where to sell

Pond5 is the specialist's choice: largest footage selection, you set your own prices, and the 30–40% royalty structure is transparent. It's competitive precisely because it's the best-known, but it's where serious footage buyers shop.

Shutterstock has enormous buyer traffic and simpler pricing, though you have less control over what your clips sell for. Adobe Stock integrates directly into Premiere Pro and After Effects, which means your footage appears in front of editors at the moment they're looking for it — a genuine distribution advantage.

Subscription platforms like Storyblocks and Artgrid offer lower per-download payouts but can move serious volume, and Artgrid in particular positions itself as a curated, higher-quality collection. Many contributors upload non-exclusively to several platforms at once — spreading the same clips across marketplaces to maximize surface area. Just check each platform's exclusivity terms before you do: exclusive content earns higher royalties on Pond5 but can't be sold elsewhere.

The technical bar

You don't need cinema gear, but you do need to clear the quality bar, and clips get rejected. The most common rejection reasons are mundane: shaky handheld footage, bad exposure, visible sensor noise, and focus issues. A modern smartphone on a tripod or gimbal, shooting in good light, clears the bar for a large portion of the market. 4K is increasingly the expectation for new uploads, though HD still sells.

Model releases are the legal detail that trips people up. If your footage shows identifiable people and you want to sell it for commercial use, you need signed releases from every person visible. Marketplaces enforce this strictly — a clip with a recognizable stranger in the background will be rejected or limited to editorial use. Shoot with friends and family who'll sign, or shoot scenes without identifiable faces.

Who this is actually good for

Stock video suits a specific temperament: patient, consistent, comfortable working without feedback. There is no audience telling you what's working. There is only a sales dashboard that stays flat for months and then starts moving. If you need quick validation, this will feel like shouting into a void.

It's also genuinely good for people who already shoot video for other reasons — wedding videographers with leftover B-roll, travelers with hard drives full of footage, drone owners who fly for fun. That footage is already sitting there. Uploading it costs nothing but the time to tag it, and every sale is found money.

The AI question: will generated video kill stock footage?

Every working contributor gets asked this, so let's answer it directly. AI video generators are getting good, and marketplaces are already flooded with AI-generated clips — which has created its own problem. Buyers in advertising and broadcast need footage they can trust: real locations, real people, model releases, editorial accuracy. An AI-generated clip of a "hospital hallway" might look fine, but no serious producer will risk a national campaign on footage that could contain subtle hallucinations — six-fingered hands, impossible architecture, a logo that almost matches a real brand.

The honest assessment: AI generation will compress the bottom of the market — generic backgrounds, abstract textures, simple animations, the clips nobody inspects closely. That's real pressure, and it means the generic-nature-shot contributor has two competitors now: a million other photographers and a machine. But it makes authentic, specific, hard-to-fake footage more valuable, not less. A real drone shot of a real neighborhood, real people signing real releases, footage of actual workplaces and events — that is what AI can't reliably produce, and it's what serious buyers pay for.

The practical takeaway: don't build a catalog of the things AI does well. Build one of the things it doesn't — genuine human moments, real places, specific activities, footage with legal documentation attached. That's the footage that survives.

The honest summary: stock video is a slow, compounding, unglamorous business. A small catalog earns almost nothing. A large catalog of searchable, useful footage earns steadily for years — one contributor's decade of uploads still paying out is the rule, not the exception. Film what's useful, tag it like a librarian, upload weekly, and don't check your earnings for the first six months. The money rewards the patient.