icantbelievethatworked: experiments that became infrastructure lessons

Not every project survives as the main thing. Some projects are valuable because they teach the next project what not to do, what to automate, and where the hidden friction lives. icantbelievethatworked belongs in that category.

Where it is now

The project exists in the active set and is also represented in archived commerce work. It sits upstream of the more disciplined FocusGoods/T-V2 approach: a place where web, commerce, and product experiments helped expose the need for better deployment, fulfilment, and operational control.

The barriers

The barrier was the classic commerce stack problem: the storefront is only the visible tip. Product data, assets, fulfilment, payment, email, rights, and customer handling all become part of the system whether you planned for them or not.

How I used it

I treated the lessons as reusable material. The later FocusGoods architecture is more serious because earlier commerce experiments made the weak points obvious. That is a real engineering habit: keep the scar tissue, but turn it into design.