Localization is not about assuming Japanese users cannot buy in English.
Some can. Many do.
I bought tea from France in French I could barely read. I subscribed to a Brazilian internet service in Portuguese I didn't understand — and kept paying until my credit card expired because I couldn't figure out how to cancel. I bought VHS tapes from US marketplaces for a film I had to watch — before I even owned a VHS player. I bought DVDs across region codes and purchased sketchy software from a Chinese site just to break the region lock. I created a dummy CPF, Brazil's individual taxpayer ID required by many checkout flows, to buy Brazilian books directly.
This was before I moved to Brazil. Before I learned Portuguese. Before any of it was easy.
I did all of this for things I already wanted badly enough.
That is the point.
Where fragile interest breaks
Most SaaS products are not evaluated from that level of desire. They are considered while interest is still fragile — "this looks useful," "maybe better than what we have," "worth trying." At that stage, English UI, unclear pricing, unnatural copy, and slow support don't make users angry.
They make users leave.
Quietly. Without feedback. Without a review. Without ever coming back.
What localization is actually for
Localization is not about assuming people can't buy in English. It is about not killing fragile interest before it has a chance to grow.