The default city break is two nights, driven by weekend availability rather than by anything about the city. For a substantial number of places it produces a rushed and unsatisfying visit.
What the first day actually costs
The arithmetic that explains why two nights is short.
Arrival day is generally half a day at best. Transport from the airport, checking in, and orienting yourself consume the rest.
Departure day is similarly reduced, since you are working backwards from a check-out time and a journey.
Which means two nights delivers roughly one full day plus two partial ones.
Three nights delivers two full days, which is a substantially larger increase than the additional night suggests.
What determines how long a city takes
Several factors that vary enormously and that people do not assess in advance.
Physical size and how spread out the things you want to see are. A compact centre is navigable on foot; a city where the attractions are an hour apart consumes days in transit.
Transport quality, which determines whether crossing the city is fifteen minutes or an hour.
Opening patterns. Cities where things close on a particular day, or close in the afternoon, or open late, compress the usable time considerably.
And density of interest. Some cities have a small number of specific things and a great deal of pleasant ambient wandering; others have more concentrated attractions than anybody can see.
The mistake of listing attractions
The planning approach that produces the worst trips.
A list of things to see, executed in sequence, produces a visit spent moving between locations and queueing.
What people generally remember from cities is not the attractions but the ordinary time — a meal, a neighbourhood, sitting somewhere.
Which suggests the useful planning unit is not a list but a rough allocation: two or three fixed things across the trip, and unstructured time around them.
A day with one thing planned works considerably better than a day with four.
The neighbourhood approach
What I have found produces better visits.
Rather than crossing the city between attractions, spending a day in one area — its market, its cafes, its smaller museums, its streets — and moving to a different area the next day.
This eliminates most of the transit, which is where a substantial share of city break time actually goes.
It also produces a sense of somewhere rather than a sequence of sights, which is the thing that makes a place memorable.
Where two nights genuinely works
Being fair, since it suits some trips.
Small compact cities where everything is walkable.
Return visits, where orientation is unnecessary and the objective is specific.
Cities reached quickly, where the travel does not consume the arrival day.
And trips with a single purpose — an event, a specific exhibition — where the city is incidental.
The timing variables
Worth checking before booking dates.
Which day museums and major attractions close, which varies by city and by institution and catches people out reliably.
Public holidays, which can close a substantial proportion of a city or transform it entirely.
Market days, which are frequently the best thing available and happen on specific days.
And whether the city empties at a particular time of year, which happens in several places and produces a very different visit.
What I would plan now
Three nights minimum for anywhere unfamiliar, four for anywhere large.
Two or three fixed commitments across the whole trip rather than per day.
Accommodation central enough to walk from, which removes transit and is worth paying for.
One neighbourhood per day rather than crossing the city repeatedly.
And an afternoon with nothing planned, which is where most of what I actually remember has come from.
The arrival day, used properly
Since it is written off in most planning and does not have to be.
A walk with no destination, in the area around the accommodation, is the best possible use of a tired half day.
It orients you, which makes every subsequent day more efficient, and it requires nothing of you.
Booking anything on the arrival evening is the common mistake, since delays make it stressful and tiredness makes it unenjoyable.
The one thing worth arranging is somewhere nearby to eat, established in advance, since arriving hungry in an unfamiliar place at nine in the evening is how arrival days go wrong.
Where to stay
The decision that determines how much of the time is usable.
Central accommodation costs more and eliminates transit, which on a short trip is the dominant consideration.
The relevant measure is walking distance to where you will actually spend time rather than distance to a nominal centre.
Proximity to a transport hub matters for arrival and departure and less for the days between.
On a two or three night trip the premium for a central location is generally recovered in usable hours, which is not true on a longer stay where the arithmetic reverses.