Ask any operator in Prenzlauer Berg which day of the year they get most wrong, and you will hear some version of the same answer. Not New Year. Not Christmas. The first properly warm Saturday.
It is worth understanding why, because the reason is not laziness and it is not bad management.
Your room doubles overnight
A café with eight tables inside and twenty on the terrace is not one business. It is two, and which one you are running on any given day is decided by the weather at about eleven in the morning.
For five months of the year you have been trading as the small version. Every instinct you have — how much bread, how many staff, how much milk — is calibrated to the small version, because that is what the last twenty Saturdays looked like. Then the temperature hits nineteen degrees, half of Berlin decides simultaneously that winter is over, and you are running the big version with the small version's prep.
Our venues consistently report this as their worst waste-and-shortage day of the year, in both directions at once: out of the three things everyone wanted by two o'clock, and throwing away the things nobody did.
Why last month's numbers cannot help you
The natural thing to do is look at what you did on recent Saturdays. That is a reasonable habit and it fails badly here, because every one of those Saturdays was cold. The average of four cold Saturdays tells you what a cold Saturday looks like. It has nothing at all to say about this one.
It fails a second time a fortnight later, in the other direction. Once you have had two or three warm weekends, they pull your average up — and then a grey, rainy Saturday arrives and you have prepped for sunshine.
The pattern is not really about temperature. It is about change. Any rule of thumb built on recent history is slow to turn, and the days that hurt are exactly the days when conditions turn faster than the rule does.
What actually predicts it
Three things, in rough order of how much they matter for a terrace venue:
Temperature over your opening hours, not the daily average. A day that peaks at 22°C at six in the evening is a very different trading day for a café closing at six than for a restaurant opening at five. Daily averages throw away exactly the part that matters.
Whether it is the first one. The first warm weekend of the year behaves differently from the fifth. There is a pent-up quality to it that fades once people have had a few. Our numbers show the first genuinely warm Saturday running well above what its temperature alone would suggest, and the effect tapering off across April.
Rain probability, more than rain. People decide whether to sit outside based on the forecast they looked at over breakfast, not on the weather that actually arrives. A dry day that was forecast to rain trades noticeably below a dry day that was not.
What to do on Monday
If you take nothing else from this: look at the ten-day forecast on a Monday in March, and mark the first day that goes above about eighteen degrees. Then order and roster for it as though it were a summer Saturday, not a March one.
That single habit is worth more than any amount of retrospective analysis. It is also, roughly, what BentoPlan is doing for you automatically — watching the forecast for your address, noticing when a day breaks the pattern of the days behind it, and telling you three days out while the order can still be changed.
The rest of the year mostly looks after itself. It is the handful of days like this one where the money is.