Utility bills have two deadlines. Almost every expense tracker is built as if they had one.

Where I live (Colombia), a utility bill prints two dates: the due date, and the shut-off date the day they actually disconnect you. Last month the gap on my electricity bill was 15 days. On my gas bill it was one.

Those mean completely different things. One costs money, the other means sitting in the dark. And the gap between them is the risk: 15 days is room, 1 day means the bill is already critical the moment it's due.

Yet almost every recurring-expense tool has a single date field. The reason is obvious once you see it they were all designed around Netflix, where one date is a complete description of the event. The whole category inherited a data model from subscriptions and applied it to bills that don't fit it.

I've been building something for my own bills, so I've hit the hard parts. Two I'm still unsure about:

• When a bill doesn't print a shut-off date, should the tool invent one? I decided no a guessed date that looks identical to a real one is worse than no date. But then some bills show less than others and it just looks broken.

• A linear timeline hides the worst case. On a real gas bill, that single day of buffer was 12 pixels out of 248. The most dangerous state was the least visible. Fixing it means distorting the scale, which is the thing you're not supposed to do.

Two questions:

Do your utility bills have a grace period between "due" and "disconnected"? And is it printed, or an unwritten thing everyone in your country just knows? A tool can't ask for a date that isn't written anywhere.

What recurring expense does your setup handle badly? Not subscriptions the awkward ones. Bimonthly bills, annual renewals you forget eleven months out, anything where the expense doesn't fit the field.