The Question Nobody Answers Straight
Ask an agency what a video ad costs and you get "it depends." Ask a creator and you get a number with no context.
Here are the actual ranges, what sits behind each one, and the hidden cost that decides whether any of them works for you.
Route 1: Creative Agency
$2,000–$15,000 per video
You get a concept, a shoot, professional editing, and revisions. For a brand film or a TV spot, this is the right call.
For performance ads, it rarely is. The economics break on volume: at $3,000 a video, a 20-creative test costs $60,000. Most brands can't run that test, so they run four creatives instead and hope one works.
Hidden cost: turnaround. Two to six weeks per batch. By the time a winner is identified, the angle may be stale.
Route 2: Freelance Creator
$150–$800 per video
You send the product, they film at home, you get raw or lightly edited footage. This is the standard UGC arrangement and it works.
The friction is operational. Sourcing, briefing, shipping product, chasing deadlines, handling reshoots — each creator is a small project. Running ten creators means running ten projects.
Hidden cost: your time. Most brands underestimate this by half.
Route 3: UGC Marketplace
$100–$400 per video
Marketplaces solve sourcing. You post a brief, creators apply, you pick. Faster than finding people yourself.
Quality varies more than the price suggests. You are buying a lottery ticket on each creator, and the usage rights are often limited to organic posting — paid usage costs extra, and brands routinely miss that clause until legal reads it.
Hidden cost: rights. Check whether paid media use is included before you scale spend behind a video.
Route 4: In-House
$500–$2,000 per video, fully loaded
Once you're producing 30+ videos a month, hiring is often cheaper per unit than outsourcing. A videographer plus basic gear covers a lot of ground.
The math only works at volume. Below roughly 20 videos a month, you're paying a salary to keep someone partially idle.
Hidden cost: fixed. Salaries don't scale down when you pause spend.
Route 5: AI Generation
$0.50–$5 per video
You supply a product photo and a script; the model produces the footage. No shoot, no shipping, no scheduling.
The cost per video is low enough that creative volume stops being a budget decision. Testing 30 angles costs less than one agency video.
Hidden cost: prompt quality. The output is only as good as the brief. Vague inputs produce generic footage, and "generic" is the one thing that reliably fails in a feed.
The Comparison That Matters
Cost per video is the wrong metric. Cost per winning creative is the right one.
Assume a 1-in-8 hit rate, which is typical for paid social:
| Route | Per video | Cost per winner | |---|---|---| | Agency | $3,000 | $24,000 | | Freelance creator | $400 | $3,200 | | Marketplace | $250 | $2,000 | | In-house | $1,000 | $8,000 | | AI generation | $2 | $16 |
The gap isn't a rounding difference. It changes what strategy is available to you.
At $24,000 per winner, you cannot afford to test — you have to be right on the first try, which means playing it safe, which means mediocre creative.
At $16 per winner, you can be wrong 30 times before lunch.
For a deeper breakdown of that trade-off, see AI UGC generator vs hiring creators.
Where AI Doesn't Win
Be honest about the limits.
Real-person trust signals. A genuine customer talking about a product they actually use carries credibility that synthetic footage doesn't replicate.
Complex demonstrations. Anything requiring precise physical interaction — assembly, texture, fit — still films better than it generates.
Brand hero content. Your homepage film and your category launch deserve a real production budget.
The practical answer for most e-commerce brands is a split: AI for testing volume, human production for the handful of concepts that prove out.
A Sensible Starting Budget
If you're spending under $10,000 a month on paid social:
- 70% of creative budget on AI-generated testing volume
- 30% held back to produce proper versions of whatever wins
Find the angle cheaply. Spend real money only on angles that have already proven they convert.