Why I Built AdLoft: The Product Photography Problem Nobody Was Solving
There’s a version of this story where I say it hit me like a lightning bolt. It didn’t.
It hit me slowly, invoice by invoice, photoshoot by photoshoot, over about 18 months of running paid ads for e-commerce brands.
The Pain Was Cumulative
The workflow was always the same:
- Client uploads product photos from their photographer
- We remove backgrounds (usually using Remove.bg)
- We position the product in Canva or Figma
- Someone — usually the client, usually the bottleneck — creates lifestyle scenes or we license stock photos
- We export the creative at the right dimensions for Facebook, Instagram, or Amazon
- We tweak again because the dimensions weren’t quite right
- Finally, we launch the ads
That’s minimum four software tools,
at least half a day of work per product, and
$200–$400 in design costs before you’ve even bought your first click.
For a brand with 20 products, that’s up to $8,000 in pre-launch creative costs — before any testing begins.
The First Version Was Embarrassing
The first version of
AdLoft
was a Python script with a terrible UI I used internally.
No product name. No landing page. Just: “upload here, get ad here.”
But it worked well enough that I didn’t want to use the old workflow anymore.
That was the signal.
When your own half-broken prototype beats the professional workflow you’ve used for years, you know you might have something real.
What We Built
The actual product generates complete ad creatives from a single product photo.
Not just background removal — full composition, contextual scene generation matched to the product category, and properly sized assets for each platform.
- Campaign Mode – clean, professional ad creatives
- Viral Mode – optimized for social content
- Rival Mode – competitive positioning ads
- Ads Mode – tuned to platform specifications
If you’re an e-commerce seller or run paid media at any scale, the math becomes obvious:
If you want to see real outputs before trusting my description, I also wrote a detailed breakdown comparing AI background removal tools vs full ad creative generation on the AdLoft blog:
What I Got Wrong
I Underestimated Quality Expectations
Early users were forgiving during testing, but brutally honest once they tried publishing real ads.
The bar for something that “looks real” in an advertisement is much higher than something that simply “looks fine on a phone screen.”
I Overengineered the UI
The first version had too many options. Users froze when faced with too many choices.
The current approach — pick a mode and generate output — converted about 3× better than the complex control panel I originally built.
I Launched Without Examples
You cannot sell an image generation product without showing real outputs.
During the first two weeks after launch, I showed a demo that was too generic. Once we added real examples from actual product categories, conversion increased immediately.
What’s Next
The product is live and generating real results for e-commerce sellers.
We’re iterating quickly on what “ad-ready” means across different platforms, product categories, and creative formats.
If you’re building something in e-commerce or AI, I’m always happy to talk.
I respond to emails.
1 Comment
Join the discussion and tell us your opinion.
My developer is trying to convince me to move to .net
from PHP. I have always disliked the idea because of the costs.
But he’s tryiong none the less. I’ve been using Movable-type on a number of
websites for about a year and am concerned about
switching to another platform. I have heard excellent things about blogengine.net.
Is there a way I can transfer all my wordpress posts
into it? Any help would be really appreciated!