how do they decide where to put bus stops?

If you are expecting consistency from something like this, you should try getting involved in local politics at least once.

Whatever way you imagine for deciding it, the answer is yes, they do that.

Do they gesstimate? Yes!

Do they hire engineers to guesstimate? Yes!

Do they concede to popular pressure? Yes!

Do they concede to money pressure? Yes!

Do they use the placement to guide the city's evolution? Yes!

I can understand the pessimism in some of the answers given so far, especially with regards to the poor state of American public transit. But ending a discussion with "they guess" is unsatisfactory to me, and doesn't get to the meat of the question, which I understand to be: what processes might be used to identify candidate bus stop locations.

And while it does often look like stops are placed by throwing darts at a map, there's at least some order and method to it. So that's what I'll try to describe, at least from the perspective of a random citizen in California that has attended open houses for my town's recently-revamped bus network.

In a lot of ways, planning bus networks is akin to engineering problems, in that there's almost never a "clean slate" to start with. It's not like Cities Skylines where the town/city is built out by a single person, and even master planned developments can't predict what human traffic patterns will be in two or three decades. Instead, planning is done with regards to: what infrastructure already exists, where people already go, and what needs aren't presently being met by transit.

Those are the big-picture factors, so we'll start with existing infrastructure. Infra is expensive and hard to retrofit. We're talking about the vehicle fleet, dedicated bus lanes, bus bulbs or curb extensions, overhead wires for trolleybuses, bus shelters, full-on BRT stops, and even the sidewalk leading up to a bus stop. If all these things need to be built out for a bus network, then that gets expensive. Instead, municipalities with some modicum of foresight will attach provisos to adjacent developments so that these things can be built at the same time in anticipation, or at least reserve the land or right-of-way for future construction. For this reason, many suburbs in the western USA will have a bulb-out for a bus to stop, even if there are no buses yet.

A bus network will try to utilize these pieces of infrastructure when they make sense. Sometimes they don't make total sense, but the alternative of building it right-sized could be an outlandish expense. For example, many towns have a central bus depot in the middle of downtown. But if suburban sprawl means that the "center of population" has moved to somewhere else, then perhaps a second bus depot elsewhere is warranted to make bus-to-bus connections. But two depots cost more to operate than one, and that money could be used to run more frequent buses instead, if they already have those vehicles and drivers. Tradeoffs, tradeoffs.

Also to consider are that buses tend to run on existing streets and roads. That alone will constrain which way the bus routes can operate, especially if there are one-way streets involved. In this case, circular loops can make sense, although patrons would need to know that they'll depart at one stop and return at another. Sometimes bus-only routes and bridges are built, ideally crossing orthogonal to the street grid to gain an edge over automobile traffic. In the worst case, buses get caught up in the same traffic as all the other automobiles, which sadly is the norm in America.

I can only briefly speak of the inter-stop spacing, but it's broadly a function of the service frequency desired, end-to-end speed, and how distributed the riders are. A commuter bus from a suburb into the core city might have lots of stops in the suburb and in the city, but zero stops in between, since the goal is to pick people up around the suburb and take them somewhere into town. For a local bus in town, the goal is to be faster than walking, so with 15 minute frequencies, stops have to be no closer than 400-800 meters or so, or else people will just walk. But too far and it's a challenge for wheelchair users who need the bus. Whereas for a bus service which is purely meant to connect between two bus depots, it would prefer to make a few more stops in between that make sense, like a mall, but maybe not if it can travel exclusively on a freeway or in dedicated bus lanes. So many things to consider.

As for existing human traffic patterns, the new innovation in the past decade or so has been to look at anonymized phone location data. Now, I'm glossing over the privacy concern of using people's coarse location data, but the large mobile carriers in the USA have always had this info, and this is a scenario where surveying people about which places they commute or travel to is imprecise, so using data collected in the background is fairly reliable. What this should hopefully show is where the "traffic centers" are (eg malls, regional parks, major employers, transit stations), how people are currently getting there (identifying travel mode based on speed, route, and time of day), and the intensity of such travel in relationship to everyone else (eg morning/evening rush hour, game days).

I mentioned surveys earlier, which while imprecise for all the places that people go to, it's quite helpful for identifying the existing hurdles that current riders face. This is the third factor, identifying unmet needs. As in, difficulties with paying the fare, transfers that are too tight, or confusing bus depot layouts. But asking existing riders will not yield a recipe for growing ridership with new riders, people who won't even consider riding the existing service, if one exists at all. Then there's the matter of planning for ridership in the future, as a form of induced demand: a housing development that is built adjacent to an active bus line is more likely to create habitual riders from day 1.

As an aside, here in California, transit operators are obliged to undergo regular analysis of how the service can be improved, using a procedure called Unmet Transit Needs. The reason for this procedure is that some state funds are earmarked for transit only, while others are marked for transit first and if no unmet needs exist, then those funds can be applied to general transport needs, often funding road maintenance.

This process is, IMO, horrifically abused to funnel more money towards road maintenance, because the bar for what constitutes an Unmet Transit Need includes a proviso that if the need is too financially burdensome to meet, they can just not do it. That's about as wishy-washy as it gets, and that's before we consider the other provisio that requires an unmet need to also satisfy an expectation of a certain minimum ridership.... which is near impossible to predict in advance for a new bus route or service. As a result, transit operators -- under pressure by road engineers to spend less -- can basically select whichever outside consultant will give them the "this unmet transit need is unreasonable" stamp of disapproval that they want. /rant

But I digress. A sensible bus route moves lots of people from places they're already at to places they want to go, ideally directly or maybe through a connection. The service needs to be reliable even if the road isn't, quick when it can be, and priced correctly to keep the lights on but maybe reduced to spur new ridership. To then build out a network of interlinking bus routes is even harder, as the network effect means people have more choices on where to go, but this adds pressure on wayfinding and fare structures. And even more involved is interconnecting a bus network to a train/tram/LRT system or an adjacent town's bus network.

When they're doing their job properly, bus routing is not at all trivial for planners, and that's before citizens are writing in with their complaints and conservatives keep trying to cut funding.

the municipality asks civil engineers, who ask their computer, then forward the results back to the municipality.

usually the public has a window where they can offer input.

if nobody checks or notices anything weird, they start allocating funding and you have your bus stops whenever they have time and funds to build them.

i like civil engineers, and a lot of them seem to care about their jobs, but their jobs have a lot of restrictions and regulations, so not everybody will be happy with the results.

Estimated usefulness compared up against practicality. Of course you want it where people are, and at yhe same time you want the bus to be able to stop without obstructing traffic.

Were I am at they have some busses that stop at the next intersection if you ring the bell and others that will only stop at designated stops. I always thought they put stops at places they predict people will want to get on and off like government buildings and major private areas like malls and then strewn ones in between at approximate set distances (every mile or two or several). Then I figure they adjust based on if people ring the bell to get off often.

If I had to guess, there are likely traffic studies done before the business get created to find most popular routes.

For existing routes, I was able to get one that was in front of my house removed because there was another one 60 feet away and not in front of a residence. I'm guessing getting one added might be a similar request.

If you need one, reach out to your local transit and ask them. Doesn't hurt to ask!

Where I live, most stops have been in place for decades if not a century at this point. No-one remembers why the all ones that exist now exist where they do, only that they exist there. Some actually migrate over time due to new construction and other factors.

But to guess how they got where they are, at least generally speaking, someone would have designed routes for public transport around main roads and important industrial areas mainly so that workers could get to work in a morning. Businesses may have even lobbied local government or bus companies for a stop near where they were if one wasn't already planned to be there.

Anecdotally, I know a stop near where my parents live was deliberately placed at the far side of a road junction so that factory workers who wanted to get off there were getting off past a fare boundary. That meant that if they caught the bus closer to work rather than a quarter mile up the road, they'd have to pay extra money. Actually, it's so old a thoroughfare it might have been a horse-drawn tram stop originally. Same fare shenanigans though.

That stop migrated to the "cheaper" side of the road junction nearly 30 years ago, but as far as I know, it's still treated as though the fare boundary occurs before it.

Anecdote 2: There have been embarrassing stories of workmen upgrading bus stop shelters only for locals to tell them, and the local news, that the bus service that would have stopped at it has long since been cancelled due to budget cuts. Bureaucracy is a wonderful thing.

It's all AI and ML using predictive models for the most efficient route planning that caters to the maximum possible population with the least carbon footprint...just kidding, it's a guy in a basement somewhere flicking his boogers at a map and deciding if those spots are good enough for his boogers then they're bloody well good enough for bus stops

Probably put them where the people are standing around in the morning.