Picture it at 2:47 in the morning: one person restyling the same plate of food for the fourth time under lights she rigged herself, on a set she built herself. Not a prop house. Not a production company with a craft-services table. Just her, a storyboard and a brand that will get one video out of the week.
That woman is Grace Wells, who posts as @gracewellsphoto to millions of followers across TikTok and Instagram, and Digiday just published her math. Roughly 100 hours per branded video. About 30 of those in pre-production alone, including a briefing call and a storyboard for every single shot. Around 20 more in the edit. The rest on set, filming, which she says is where the bulk of her time goes. Two to three branded videos a month. That’s her whole capacity, and brands are filling it.
Read that again in the context of every pitch deck you’ve sat through this year.
The industry spent 2026 selling the opposite promise. Fifty variants per shoot. A hundred cutdowns per campaign. Talent cloned so one day on set becomes a year of assets. The same week Digiday ran Wells’s numbers, it reported that Ogilvy and CAA had launched a partnership to build AI versions of celebrity talent for ads, with the partnership expecting cost per asset to fall 65 to 75 percent. That’s a real business with a real logic. We’re not here to pretend it isn’t.
We’re here to point out that it’s a different business.
For about fifteen years, content economics ran on a single assumption: more output beats better output, because the feed is infinite and attention is cheap. Every brand became a publisher. Every publisher became a factory. The metric was volume, and the job of the agency was to make volume look intentional.
That assumption is now splitting in two, and most marketing budgets haven’t noticed. On one side sits the volume market, where machines are cutting the cost of a usable asset toward zero and every competitor has access to the same machines. On the other side sits the craft market, where a single person spends a hundred hours on one video and somebody pays for it, every month, on purpose.
Here’s the thing about a market where everyone has the same machine. The price of what the machine makes falls to what the machine costs. That’s not a hot take. That’s what happens to every commodity, from stock photography to wedding DJs with Spotify. If your content strategy is “produce more than the other brand,” you’ve just entered a race whose finish line is free.
Craft doesn’t work that way, because craft is the one input the machine can’t make cheaper. Thirty hours of pre-production is thirty hours of somebody deciding what doesn’t go in the frame. You can’t prompt your way to that. You can only spend it.
And audiences can tell. Not because they’re film students, but because they’ve been trained by a decade of feeds to clock effort in under a second. The thumb stops for the shot that obviously cost somebody something. The thumb keeps moving for the shot that looks like the last forty shots. Digiday’s own reporting cites YouGov finding that 49 percent of Britons view AI-generated actors in ads negatively, and a DoubleVerify and YouGov study in which 92 percent said AI-generated content makes it harder to trust what they see online. That’s not a niche of purists. That’s the room.
Wells, for her part, is blunt about where the time goes. “I’m very big on pre-production,” she told Digiday. That sentence should be printed on the wall of every content team that skips the storyboard because the edit is “where the magic happens.” The magic happens when somebody decides what the video is before the camera rolls. Everything after that is execution.
Now the turn, because you’re already composing the reply.
“Sure, a hundred hours works for one creator doing two videos a month. It doesn’t scale. We have twelve markets and a product drop every Thursday.”
Correct. It doesn’t scale. That’s the point. That’s the entire point.
The reason a hundred-hour video is valuable is precisely that your competitor can’t produce forty of them by Friday. Scarcity is the product. The minute craft scales, it becomes volume, and you’re back in the race to free. So the question isn’t how to make craft cheaper. The question is how to decide which moments deserve it, and how to stop spending craft money on volume work and volume money on moments that needed craft.
Most brands get that allocation exactly backwards. They pour production budget into the always-on grid, the forty identical carousels a month that nobody remembers, and then hand the launch film, the one asset that will actually carry the brand for a year, to whoever had a free Tuesday. The result is a feed that’s too polished to feel human and a hero piece that’s too rushed to feel like anything.
Flip it. Let the machines handle the grid. Let them resize, translate, cut down and test, because that work genuinely benefits from speed and nobody ever fell in love with a resize. Then take the hours you saved and spend them somewhere they compound. One piece a month that someone built by hand. One shoot where the set was designed, not rented. One idea that took thirty hours to find before anyone touched a camera.
There’s another shift hiding inside Digiday’s reporting that makes this sharper. Brands are rewriting creator rate cards around performance, asking for cost per lead, cost per conversion and earned media value before they ask about follower counts. On the surface, that sounds like the death of craft: show me the numbers or get out. Look closer and it’s the opposite. When you pay on outcomes, the creator who makes something people actually stop for wins, and the one who floods the feed with filler loses. Performance pricing is the market finally pricing care.
We live in a city that understands this better than most. Los Angeles is full of people who spend a week on a single shot, from the grip crews in Burbank to the muralists in the Arts District who rework a wall until it’s right before anyone sees it. Nobody calls that inefficient. They call it the work. Somewhere along the way, brand marketing convinced itself that its work was different, that a post was a unit and a unit was a cost and the cost should approach zero. That’s how you end up with a feed full of things nobody made.
We run our own shop on the other side of that line, and we’ll admit it’s not always comfortable. Craft is slower to sell. It’s harder to put in a spreadsheet. It forces a conversation about what matters, which is a conversation a lot of marketing teams are built to avoid. But it’s the only conversation that ends in work people remember.
So here’s the uncomfortable audit. Pull up the last thirty days of your brand’s content. Count how many pieces somebody spent more than ten hours deciding on before production started. If the answer is zero, you don’t have a content strategy. You have a content habit, and habits are exactly what the machines are about to do better and cheaper than you.
Grace Wells isn’t slow. She’s one of the few people in this business still doing the expensive part.
A hundred hours isn’t the cost of the video. It’s the reason anyone watches it.
