It usually starts in the dark.
The atelier lights are off. Nils Frahm is on the speakers — the slow piano stuff, the records where you can hear the felt of the hammers. I'm sitting with my reporter's pad open, jotting down ingredient interactions I want to chase. A note about how a particular Elemi might behave next to a synthetic muguet. A reminder to revisit an oud profile I half-abandoned last winter. The kind of fragmentary scribble that means nothing to anyone else and everything to me three weeks from now.
This is the front end of every perfume I make. Quiet. Slow. Mostly internal. It looks more like a writer working through a difficult chapter than anyone's idea of perfumery.
Then, once the concept has enough weight to start pulling materials, the room flips. Lights up full. Playlist swings to British new wave — Magazine, Wire, the louder end of Echo & the Bunnymen. I'm up, moving, opening bottles, weighing, smelling, scribbling, walking circles. It probably looks manic. It is, a little.
Most of what begins this way never leaves the lab.
The math, since people always ask
A few real numbers, because I think it helps to demystify this.
la Un.e — our most worn perfume, the one that quietly anchors the whole house — came together in four iterations. Four. It's still surprising to me. Some formulas arrive almost fully formed, like they were already in the room before I walked in.
Bois d'Agar 01 took thirteen. That feels closer to the average for something built around a serious natural.
The more complicated limited runs — the ones where I'm trying to do something I haven't done before, or where I'm balancing several rare materials against each other — those will run twenty, sometimes thirty iterations before I either ship them or pull the plug. Anything past thirty and I'm usually fooling myself.
The harder number to give you: more than half of the perfumes I actively start formulating for Les Vides Anges never make it into a bottle. Some die fast, in a week. Others I let sit for months, occasionally up to two years, while I figure out whether the material wants more time or whether I just don't have the answer for it.
Rough edges are not the same as failure
I want to be careful here, because I think these get confused.
I like rough edges in a perfume. I leave them in on purpose, particularly in the small batches where I have a rare natural doing most of the work. Sanding everything smooth is one of the things that makes commercial fragrance smell commercial. There's a softness, a roundness, an inoffensiveness that gets engineered in at the late stages of development, and to my nose it tends to flatten whatever was interesting about the brief in the first place.
A rough edge means the formula is finished. I just chose not to polish it.
Failure is a different thing entirely. Failure is structural. It's the heart notes refusing to sit where I want them. It's a base that swallows everything else in the formula and starts narrating the whole perfume in its own voice. It's a top that smells synthetic in a way that has nothing to do with whether the materials are synthetic. It's a formula that smells fine on a blotter and dies on skin, or one that opens beautifully and then ten minutes in turns into something I don't recognize and don't want to know better.
Rough edges are a choice. Failure is the formula telling me it isn't going to work, and I'd rather hear that early than fight it for another six months out of ego.
The blind spots I'm honest about
There are entire categories of perfume I am unlikely to ever release under Les Vides Anges, and they are exactly the categories I formulate most often for clients.
Amber-heavy bases. Big plush oriental territory. Soft, sweet, ambroxan-driven backbones with resinous warmth piled on top. That's the bread and butter of commercial reformulations and a lot of new-brand briefs, so I work in that space constantly. But when I try to bring it home to LVA, something in me checks out. I lose patience with the formula faster than I should. I stop being interested in what it could become.
Citrus-forward perfumes are the same way. I'm not going to create something in the lineage of Acqua di Parma, and I'm not going to create the inverted version of that — the dense, animalic, almost confrontational Orto Parisi or Nasomatto end of things. Both extremes give me difficulty when I'm formulating for LVA. I respect them. I don't want to make them.
So when a formula in either of those territories starts to wobble — when a heart note won't lock in, when the base starts overplaying its hand — I think I let go faster than I would for something I'm actually compelled by. Is that self-sabotage? Probably, a little. I've made peace with it. There are only so many hours in a lab year, and I'd rather give them to the perfumes I actually want to wear.
This is one of the realer answers to why so many formulas die: not because they were technically broken, but because I was the wrong perfumer for them.
The technical kills
When something does die for craft reasons rather than taste reasons, it's almost always the heart.
Les Vides Anges perfumes don't lean on top notes. I rarely weave more than a sketch of one in. The opening of an LVA perfume is short on purpose — I want you in the heart within minutes, because that's where the perfume lives. Which means the heart has to carry an unusual amount of structural weight, and when it won't, I notice fast.
A heart note that won't sit is a specific feeling. The formula keeps wanting to revert. You'll add a counterweight, smell it the next week, and the offending material has somehow reasserted itself anyway, like a piece of furniture that keeps drifting back to the middle of the room, no matter how many times you move it. After a few cycles of that, you know.
The discipline is in knowing the difference between won't sit and hasn't sat yet. Some heart materials take real time. The pineapple aromatic family — Allyl Amyl Glycolate, Allyl Heptanoate, the molecules behind a lot of what people associate with the Creed house — are powerful but unwieldy when you first put them into a formula. They dominate. They sit on top of everything else like a bright loud guest at a quiet dinner party. If you panic at week two and pull the formula, you'll never get to see what those materials do at week six, when they've integrated and started to play nicely with what's around them.
So I let formulas sit. Weeks. Months. Sometimes I'll come back to a trial I'd written off and find that the perfume has done the work without me, and what looked like a failure was just impatience.
Other times I come back and it still smells wrong, and I drop it.
What I save, and what becomes something else
Nothing is ever really lost.
A formula I've abandoned for LVA might quietly turn into the spine of a client brief. A heart accord I couldn't make work in one context might be exactly what someone else's project needs. I strip components out, bank them, label them, and they sit in the archive until they're useful.
Some of the best material in the Les Vides Anges line came directly out of so-called failures.
The fresh-air accord at the heart of Eau de Surréel is the cleanest example. It started as an experimental mistake. I was working on something else entirely — I don't even want to admit which formula — and noticed how the Elemi natural we use was interacting with a particular synthetic muguet I had on the bench. Biomuguet, to be specific. Something happened in that combination that I hadn't planned for and couldn't have predicted from looking at either material alone. It was wider, more vertical, more outside than either ingredient suggested on its own.
The formula I was actually working on died. The accord survived, got rebuilt around itself, and became the reason Surréel exists. Textbook happy accident.
I try not to romanticize this. Most failed trials are just failed trials — the materials don't have a hidden second life. But you learn over time to keep one ear open during a kill, because every so often the thing that wasn't working contains something that is.
The cost, which is real
I should say something about money, because I think people underestimate how expensive this kind of work is.
The agarwood natural we use across the Bois d'Agar line costs more per ounce than gold. That isn't a flourish — it's the literal pricing. Luckily a little is all you need, but a single series of experiments around a rare natural can easily run into the hundreds of dollars of pure material cost before you've made anything releasable. Multiply that across the half-or-more of formulas that never ship, and you start to see why hand-mixed limited-run perfumery doesn't scale the way a mass-market line does, and why the bottles cost what they cost.
I'm not complaining. This is the model I chose. But it's the reason I'm diligent — sometimes painfully diligent — about how I run experiments with the rare materials. Wasting an ounce of the wrong synthetic stings a little. Wasting an ounce of the right natural is something else.
What's actually in the bottle
When someone holds a Les Vides Anges perfume and sprays it, all of this collapses into one impression. They smell a single thing. The work isn't visible.
But the bottle is the survivor of a much larger process. There are formulas that came before it and didn't make it. There are materials that fought each other and had to be separated. There are happy accidents, like the one that gave Surréel its shape, that came out of something else's collapse. The perfume in your hand isn't just the work of the formula that succeeded — it's the work of all the formulas that taught it how to.
Aroma chemicals and naturals have personalities. They want what they want. Some are sledgehammers and some are feathers, and a lot of the craft is in knowing which is which at any given moment in a build, and adjusting your hand accordingly. Sometimes you're trying to coax. Sometimes you're trying to restrain. The materials are not neutral and they are not infinitely malleable, and pretending otherwise is how you end up with a formula that fights you for two years.
But that's where the fun lies, for me. The knowledge. The deep research. The experimentation. The time and patience to make something work that doesn't want to. The puzzle of it.
Failure is the cost of being interested in the puzzle. More than half of what I start never leaves the lab. The half that does is what shows up in your bottle, and I'd rather it arrive that way — chosen, finished, occasionally rough at the edges — than smoothed into something neither of us would remember.
— August
