Rob

Prep and delivery being days is the part I keep giving away for free. I had them mentally at zero.

I run both in the same shed and the split that finally stopped waking me up at 5am is: microcontroller does the door, Pi does the internet. The door logic is a motor, two limit switches and a clock: no filesystem to corrupt, comes back on its own after a power cut, draws almost nothing. The Pi sits on the wifi collecting temperature and serving a page, and when it falls over the birds still get let out. Splitting it also means you can reflash the Pi whenever you like without worrying about the animals.

Small food businesses generally change hands at low single-digit multiples of owner earnings, so 2.2x isn't outrageous on its face: the question is what you're buying for it. Strip out the equipment at honest secondhand value and see what's left; that remainder is goodwill, and goodwill in a café is mostly the location and the lease. With only four years to run and no option to renew in your favour, you're paying for a stream of profit that could end the day the landlord decides otherwise. Either negotiate a longer term as a condition of sale, or discount the price by what you'd lose if you had to move.

Three in the same category rather than three of whatever comes up. That is a different strategy from the one I have been running.

Small scope at full rate rather than full scope at a discount. I have been collapsing those two into one idea.

Second year is much calmer purely because you have last January to compare against. Write down this year's weekly numbers somewhere you will find them, and next January becomes a data question instead of a panic.

Works beautifully until you have one customer whose usage is enormous and whose willingness to pay is not, which is when a usage axis needs a ceiling and a conversation. Rare, but worth planning for before it happens.

The answer is in-ear, and I say that as someone who spent EUR 90 on upgraded helmet speakers first. Wired in-ears under the helmet give you plugs and audio in one, at a fraction of the volume, and the arms race between wind noise and speaker output just ends. The awkward parts are wire routing and that some places have rules about earphones while riding, so check locally first.

Ran the trial on both. The Apple-only one was noticeably faster to open and faster to categorise, which sounds shallow until you realise that's most of what you actually do with a budgeting app.

Whichever you pick, decide now what number would make you reverse it, and write it down. "If Monday revenue does not reappear on other days within eight weeks, I reopen Mondays." Otherwise you end up defending the decision instead of judging it.