San Francisco to Start Smart Parking Management Trial Soon

Using the new credit-card enabled parking meters. Photo: ##

The central principle of San Francisco’s cutting-edge parking management program, SFPark, comes right from Econ 101. If there are more people looking for parking than there are parking spaces (i.e. demand is greater than supply) adjust the price of parking until there is enough turnover on a given street, or roughly one free parking space per block. Sounds simple in theory, right?

On the other hand, implementing the principle in real-world conditions at over 6,000 curbside parking spaces and 11,500 off-street spaces in city-owned garages is very complicated. The federal government, which has paid for most of the program with approximately $20 million in grants, wants proof that San Francisco can meet its stated goals of reducing traffic and speeding up transit with smart parking management. That will require copious data and extensive analysis.

Most importantly for parking managers at the San Francisco Municipal Transportation Agency (SFMTA), they want the public to like it. If a driver doesn’t get to a parking space quickly, thus reducing the cruising for spaces that generates up to 40 percent of local traffic in some cities, then the program won’t deliver on its goal. Similarly if drivers aren’t happy with the convenience of the new meters or other payment options, like pay-by-phone.

Jay Primus, SFPark’s manager, understands the significance of his work and has been spending most of his waking hours for the last three years at work or conducting outreach with businesses, politicians and community groups.

As I sat down recently with Primus in his windowless office on the top floor of the SFMTA building at 1 South Van Ness Avenue, I was impressed with the impeccable order he kept. Two rows of more than twenty manila envelopes were lined neatly on a table near his desk, each representing a different part of the project, from a folder bearing the name of the communications consultants he hired, to another for grant obligation deadlines to the US Department of Transportation.

Several enormous maps of San Francisco and the SFPark areas adorned his office walls. One map, approximately 5 feet by 5 feet, showed every publicly available parking space in San Francisco and represented the first census of parking spaces conducted in any city in the world. On another wall, Primus had a scrolling work-flow plan tacked right above a cover story in the tabloid San Francisco Examiner with the inflammatory headline, “Parking Privileges to Be Revoked!”

Primus, a tall man with a studious mien and a quiet voice, worked as a transportation planner for a private firm before joining the SFMTA in 2007 to direct SFPark. He measures most of his words carefully, often stopping mid-sentence to replace technical jargon with more pedestrian language.

Primus tells me the public’s reaction to the new meters that accept credit cards has been “largely positive.” He acknowledges that some have complained that the meters are not easy to read at night, but he says increasing the back-lighting uses more power and shortens battery life.

In addition to the information the SFMTA will gather from the new parking meters about how people choose to pay and how long they pay, the agency has installed occupancy sensors in the pavement in SFPark areas that provide real-time information on how long cars are parking at spaces. When I ask him about the information the agency is already collecting, his eyes light up. “One of the most exciting things about SFpark is the fabulous, unprecedented data set,” he said. “I do believe it’s the first of its kind.”

The Unprecedented Data Set

Primus and his team will cross-reference those parking data from the meters and sensors with citation data; travel demand data from regional and city roadway sensors; transit boardings from BART and Muni; parking tax and sales tax returns; collision statistics from the SFPD and the state; manual data collection such as driver intercept surveys, parking search time surveys, double parking counts, disabled placard use, occupancy information in residential neighborhoods adjacent to SFPark areas; and exogenous statistics like the cost of gasoline, the unemployment rate, the consumer price index and hourly precipitation.

Comparing parking data to transit boardings on Muni is not trivial, Primus explains, because he will have to demonstrate the effect smarter parking management has on transit travel time and transit delays. If there are open spaces at the curb, in theory there should less double-parking and fewer delays to buses. Measuring parking and sales tax returns or gas prices should let the SFMTA know how much of the reduction in traffic is due to SFPark and how much is due to larger economic patterns. He even hopes to show that better parking management reduces traffic collisions and increases safety as drivers cruise less for an elusive space.

The obvious implication about the status quo in San Francisco and every other city that doesn’t collect this information is that policy makers know embarrassingly little about how the standards, prices and regulations they put on parking actually effect traffic, the way people park, or even how people feel about parking (and in civic life, parking is almost as emotive an issue as the crime rate).

Donald Shoup, the UCLA economics professor and author of The High Cost of Free Parking, whose work is the theoretical underpinning of SFPark, has a favorite adage that he quoted to Streetsblog when critiquing how parking policies are currently set nearly everywhere: “You can’t manage what you can’t measure.”

SFPark should give San Francisco managers an unparalleled road map whereby they can make educated policy decisions and they can measure the impact those have in real-time.

Improving the Public Perception of Parking

“We’ve done a lot of outreach with SFPark and have talked to a lot of community groups,” said Primus. “People we talk to are sometimes skeptical about the SFMTA’s intentions, that somehow SFPark is meant to gouge drivers for additional parking revenue. That’s just not the case.”

He is actually optimistic the program will become popular with drivers, for numerous reasons. “We hope to earn people’s trust that SFMTA’s parking management can help achieve our goals for the city,” he said.

Click to enlarge this image of the SFPark iPhone App, which will be available soon in the iTunes Store. Image: SFMTA

In addition to the convenience of paying at the meter with credit cards and extending meter limits up to four hours in certain areas, when the SFMTA officially launches the program later this spring, it will provide a map with real-time occupancy and price rates at every meter and parking garage in SFPark. The SFMTA will also release an iPhone app at the launch, with other app formats to follow.

The maps will color-code blocks, with dark blue indicating there are available parking spaces, light blue showing fewer spaces, and red suggesting drivers park elsewhere. Each public garage in the program will be indicated with a large P icon and will be represented with the same color scheme.

Just as drivers look at real-time traffic information on Google Maps, for instance, Primus imagines drivers will check for parking availability at their destination even before they get in the car. What about those drivers who would check their phones while driving? Primus explains that the app uses the phone’s GPS and has an automatic warning if it detects the phone is moving faster than 10 miles per hour. This feature, said Primus, will “remind people that it is illegal to use a cell phone while driving.”

The map and the mobile app will also display pricing information in various shades of green (“for money,” said Primus), with details on parking rates by hour and by location at the tip or one’s fingertips. All real-time data will be made available on an open API for third-party developers as well.

One thing the SFMTA won’t do is give availability by individual parking space, though they have that specificity internally. “We don’t’ want people to race to get to a space or fight over spaces they feel ownership of,” said Primus.

Perhaps the most important part of improving the convenience of parking with SFPark is an option to pay by phone at meters and garages throughout the city. Though the service won’t be activated until later this year, and still needs SFMTA board of directors approval, Primus said they intend to offer the pay-by-phone service at every one of the nearly 27,000 meters citywide.

This would allow anyone who has signed up with a credit card to pay for parking through their phones, to get updates automatically to their phone when time is running out, and to pay for more parking with the touch of a button, so long as they aren’t exceeding time limits. No more leaving a restaurant or a business meeting to feed the meter, said Primus. A similar service is already operational in over 100 cities throughout North America and Europe.

Long-Term Impacts of the Trial

By law, the SFPark meter rates will only change once a month at the most, so drivers shouldn’t expect a price shock. Nor will the rates likely change that dramatically, according to Primus. In some cases, where there are many vacant spaces in an area, the meter rates could come down.

No one will know how it all works before the trial starts, but the SFMTA expects to gain efficiencies in meter maintenance and enforcement. As Primus noted, the meters will instantly communicate with his database when they go out of service, so meter technicians won’t have to guess or do broad sweeps to find malfunctioning meters.

Enforcement will be much more precise as well, though Primus doesn’t expect to see ticket blitzes. Rather, he argued, with longer time limits and easier ways to pay, such as pay-by-phone, he thinks PCOs will write fewer tickets for meter violations. “We want PCOs to have more time available to enforce other issues, such as double parking, sidewalk parking or driveway parking, issues that effect transportation, quality of life and access more generally,” he said.

In fact, Primus expects meter-related citations to drop significantly as people find it easier to pay. Rather than the current “punitive” ratio of $34 million in meter revenue and $30 million in meter-related fine revenue each year, Primus hopes to see most of the revenue coming from proper payment. “People pay for parking one way or another, either at the meter or with parking tickets. For everyone’s benefit we want everyone to pay at the meter to reduce the number of parking-related tickets we have to give,” he said.

Donald Shoup didn’t mask his excitement about the impending start of SFPark, which he characterized as the most significant example of parking reform to come in the six years since he published his 750 page epic (the book has been so popular it will be released in paperback later this year). Shoup said “academics are just drooling about all this data” and he predicted legions of PhD dissertations to result from SFPark.

Most importantly for cities, though, he hoped to see a direct relationship in the data between parking and economic activity. Good parking management “can make the whole transportation system perform better because there is less cruising and it will make the whole economy perform better,” he said. “If this relationship appears in the data, it will show people this is a very powerful tool for economic development in cities.”

  • Nick

    How exactly will a sole driver access this ap without violating the driving while texting law?

    Double park in the bike lane?
    Block a crosswalk or fire hydrant?

  • icarus12

    Why not allow parking for four or more hours at nearly all meters? As long as one is paying the market price for street parking, what does it matter how long somebody parks? Then you really would be collecting nearly all your revenue from pay-to-park, instead of slapping tickets on the person who is trying to run back to the 1 hr meter to feed it a couple of more quarters.

  • TK

    Sigh. You know, Nick, this question’s been hashed and re-hashed in places like SFGate, and really, I think responsible adults can figure out the answer to that question. In fact, if you re-read the article, you will see that they even over-engineered the app and will pop up an annoying reminder that it’s illegal to use phones while driving, if the thing detects you looking while the car is moving. I’m picturing myself getting mighty sick of that when I try to find my husband find a parking space with this thing!

    Anyway, I don’t know, hire a buddy to ride with you to look at your phone?

  • TK

    er, “*help* my husband find a parking space….” etc.

    Oh, and I was going to follow up on Icarus’s question. What’s really irritating is getting a ticket for exceeding a meter even though it’s been paid up just because it’s got a 1 hour limit–on a not-so-busy commercial street (I’m looking at you, Ocean Ave.)

    Is there some way we can find out which streets are getting the new meters on what timetable, and if there’s a way to petition MTA to extend meter limits on certain blocks? On the south side of Ocean Ave east of 19th, for example, there are a bunch of doctors’ offices. It’s absurd that one can only park for an hour, when doctors (and dentists, for sure) might keep you waiting for a while. What are they going to do, have a nurse’s assistant move your car for you? Instead, we just have to piss off the neighborhood and park in front of their houses (or take Muni, but that’s not feasible for everyone, all the time).

  • Mario Tanev


    The meter price may vary according to time of day. If a driver is parked during periods with two different prices, what total price should they be charged? As you can imagine, it can cause a lot of confusion as drivers will demand they know up front what rates they will be charged. The simples way to do this, is to simply charge the rate at arrival. But if all-day parking is allowed, then some will be getting a very good deal if they park at 4 a.m. and hog up the space.

    A phone app can perhaps solve this. Since rates are not truly dynamic (they are recomputed once a month), then the phone app can perhaps inform the driver how long it will cost if they stay for 1, 2, 3, 4, etc. hours. If rates were truly dynamic that wouldn’t be possible, though (and dynamic rates are the best way to deal with one-off phenomena, like festivals, or special events).

    One potential way to deal with this while allowing all-day parking price increase depending on duration of stay (perhaps exponential, approaching the price of a ticket).

  • Disco Burritos

    @nick, by double parking, but in theory for less time than one might spend circling the block looking for parking? or at stop lights, maybe.

    @the author,

    “He even hopes to show that better parking management reduces traffic collisions and increases safety as drivers cruise less for an elusive space.”

    The way this is phrased suggests that the studies may be manipulated to show the desired result. I hope this phrasing is just a product of Streetsblog’s reporting, and that Primus genuinely wants to test the hypothesis, rather than find a way to prove it with data.

  • @Disco Burritos,
    I think you’re reading too much into “to show.”

    Try this sentence:
    “He even hopes that better parking management reduces traffic collisions and increases safety as drivers cruise less for an elusive space.”

    If it does that, he’ll have the data that will show the US DOT that it is the case.

    That’s what I meant by “to show,” but I see your point with the phrasing.


  • EL

    According to Primus: “In some cases, where there are many vacant spaces in an area, the meter rates could come down.”

    Shouldn’t that be “will”, not “could”? Isn’t that the point of demand responsive pricing? After all, Matthew Roth wrote: “If there are more people looking for parking than there are parking spaces (i.e. demand is greater than supply) adjust the price of parking until there is enough turnover on a given street, or roughly one free parking space per block. Sounds simple in theory, right?”

    The fact of the matter is that if the rate doesn’t go down in underutilized areas, then SFpark would look a lot like a price gouging scheme – much like how Redwood City (another Shoup model) is currently trying to raise rates even though the occupancy levels, when the meters are running, don’t justify it.

  • ramona rey

    I’ve stopped going to neighborhoods where it is impossible to park. I believe in public transportation but not at the expense of safety. More public parking lots? Goodbye North Beach and Mission.

  • Otto


    The only areas I don’t usually expect to be able to park are Downtown, North Beach, ChinaTown and Telegraph, Russian and Nob Hills. Those areas, I take transit. Otherwise I can generally always find a parking place, metered or not.

    The Mission is actually pretty easy.

    As for whether SFPark will work, we’ll just have to see. But there’s a part of me that thinks this system is too smart for it’s own good.

  • Alai

    @Mario: I don’t see the problem. The rates should be displayed – $3/hour 7-9AM, $2/hour 9AM-12PM, etc. Someone who arrives at 8 AM and stays two hours should be charged $3 + $2. The display could show something like “$3/hour” in big text, then “$2 per hour after 9AM” in smaller text.

    @Ramona: I can’t tell if you support this project or oppose it.

    @icarus12: I agree.

    I support this project in general. The biggest pitfall I foresee is cars with placards. Suppose the street is popular, and the project is a success: prices go up but there are ~15% vacancies. Great. But now drivers with placards will know that too, and won’t care about the price, so they’ll head straight to it and fill up those spaces. So, no vacancies. So the price goes up, and up, which succeeds in freeing up spots from ordinary people, but the spots are filled with placarded cars whose drivers don’t care about the price. It ends up being a de-facto handicapped placard zone with a ridiculous price for everyone else, annoying everyone, and is declared a failure and the project is canceled.

    The only way to succeed, I think, is to charge people with handicapped placards. You can give them a 50% discount, or allow them to exceed ordinary time limits, but without some disincentive to park there I fear this won’t work.

  • EL

    I’d have to agree with you Alai. The placards may very well doom SFpark. There’s no way to implement a demand responsive pricing scheme when so many don’t even have to pay. According to sfgate, 52600 placards were issued in San Francisco County alone by the end of 2009, and you know that number has only gone up.

    What’s even worse is that as the prices go up, it only increases the cost effectiveness of a placard, and hence (following basic economic theory) will make the abuse even worse.

  • “Finding parking in downtown SF is hard.”


    “Finding *free* parking in downtown SF is hard.”

    Okay, you make a good point. OTOH, parking in one of the many available garages is easy and convenient.

    If you’re really such a skinflint, try ditching your car. You’ll save thousands of dollars per year, not to mention the cash windfall from the sale itself.

  • TK

    I’m sorry if the information’s here in the article, but I’m still hoping someone will answer this: Does a new meter in a space automatically mean the total allowable park time is extended, i.e. no more 1-hour-limit spaces? And are these things being installed citywide on an aggressive schedule?

    I think others are quite confused about the pricing scheme. I’m fairly positive there will be no price changes within a single day, for example. In fact, I believe the article says prices won’t be changed more than once a month, by law.

    I also think people are confused about the distorting effect disabled placards have. If you have no way to get around town other than a car, I think your demand for a space is pretty inelastic. If anything, it would help everyone for DP holders to know good times to go downtown to run errands or whatever by checking the parking availability app. Isn’t that kind of the point, to move more parkers to off-peak hours?

  • EL


    I believe that when a new meter is installed, they generally match the time limits of the surrounding area. I think 1-hour limit spaces are stupid.

    If you follow the Shoup model, the price will vary throughout the day, and the price could also change each day. In fact, the SFpark website says that some time periods will cost more than $6 an hour for “special event” pricing. See here:

    The only thing SFpark commits is that once the day/hourly rates are set, it won’t change for a minimum of 1 month.

    There is no confusion regarding the effect disabled placards have. Go to downtown or Chinatown on any weekday and look at a parking space that isn’t a loading zone. Probably at least 8 out of 10 times you’ll see a disabled placard, which allows you to park for free. SFpark changes none of that.

  • Donald Shoup, who has written the best book on the topic, says that there should be no time limits on parking. They should just use pricing to get the vacancy rate they want.

    I don’t see a problem with variable pricing. Just charge them the different prices for the different times that they are there. They would have to pay electronically to make this possible.

  • Can I get an app to tell me what the heck the text on the meter says after dark? On one of my rare driving occasions, I parked at an SFPark meter and had a devil of a time knowing what buttons to push on the parking meter because I couldn’t read the little LCD screen in the dark. Perhaps some of the additional revenues could be used for back lights or some kinda lighting so drivers can read the instructions?

    Happy as can be that demand pricing is going to kick in … too bad handicap placards still trump it though.

  • Alai

    I have to comment on this: “If you have no way to get around town other than a car, I think your demand for a space is pretty inelastic. If anything, it would help everyone for DP holders to know good times to go downtown to run errands or whatever by checking the parking availability app. Isn’t that kind of the point, to move more parkers to off-peak hours?”

    The effect would be quite the opposite. SFPark aims to make parking available at all times, by increasing the price at times of peak demand to encourage turnover. Currently DP holders may be able to park for ages and for free, but they still have to hunt for spots along with everyone else.

    With SFPark, the idea is to ensure the availability of empty spots by adjusting the price. But DP parkers will happily park at these spots regardless of the price, so the price-adjustment mechanism will be unable to manage demand, and the price will increase drastically. You’re back to square one, only at a much higher price. The problem is especially acute when SFPark is only a trial, because that will make those limited areas extremely attractive to anyone with a placard.

    The only way for it to succeed under the current rules is for the supply of parking spots to significantly exceed demand from all the DP holders, but that’s effectively impossible when you’re dealing with 50,000 of them.

  • Disco Burritos

    @Alai, you keep quoting this 50,000 number, but what is that relative to the non DP drivers in the city everyday? I suspect you may be overestimating the effect the DP drivers may have, but we’ll have to wait to see what the data tells us. Hopefully meter maids will be documenting when this occurs.

  • Alai

    Sorry, I was just quoting someone else upthread. Here’s a reference, though, which gives 49772 in 2006: gives 385000 vehicles ‘in San Francisco’ households, to which you can add work vehicles and visitors.

    Still, over one in eight local vehicles is a lot, and the people with placards likely use their cars more because of the enormous advantage in parking, so the anecdotes about half or more of the cars on congested streets sitting all day with handicapped tags don’t seem unrealistic. Don’t know if anyone’s done a good study.

  • EL

    Alai, see this article from 2008. The Controller’s office determined back in 2007 that there’s $15 MILLION in lost revenue due to placards. The last paragraph in the article mentions some usage observations done by the Examiner.

    Again, this article is from 2008, and the price at the meters has only gone up.

  • olen

    OMG. What a hassle! Have humans really made their lives so difficult as to warrant such complexities? It’s astounding how much effort is going into maintaining such an unhealthy habit for such a tiny, dense city. Why would anyone bother keeping a car in SF unless they are mobility-impaired? I’m holding out for teletransport, or at least public-use Google robot taxis that don’t need to park, ever.

  • Hlamb70488

    Placard and Handicapped parking is a sham.  I did two studies for my Masters and found that 9 out of 10 handicapped vehicles were actually drive by non-handicapped persons. The handicapped sticker belonged to grandma who was at home in bed.  Several were dead and several others were in nursing homes.  I did not try to do a whole city, but chose 20 city blocks at random and began my survey.  The Placard group was 6 out of 10 for real use, electricians, plumbers, internet and phone etc.  The other 4 were bogus, electrician getting groceries with wife along, AT&T truck with man and wife at liquor store.  4 were police person, out of uniform buying groceries, cigarettes and liquor.  Lawyers are more prone to just pull up and park anywhere and resolve the ticket later with either the judge or the cop.  I saw many cops, meter maids, just ignore the lawyers car as if it was not really there.  Gave the car in front a ticket and behind a ticket.  Each thought they could park for free because the car present had no ticket and with an expired meter.  Ha ha.  Seeing is believing if this will solve anything.  I say charge everybody all the time.  Let the Handicapped be reimbursed by their insurance or doctor etc.  Let the utility company add the cost of parking be added to cost of the contract.  If the work is for the city, they will collect it back on the meter and if its private, the business will pay the contractor.  Everybody pays.  The one hour limit will work wonders for Doctor offices.  Patients will run out to move their cars and have to drive around the block seven times looking for a new slot.  Where did this idiot come from thinking he can macro manage every bitsy piece of our lives, just let the price escalate by 20% after the first hour and another 20% after the second hour, at least the patient doesn’t have to run out of the doctor’s office in a dressing gown to move their car.


Census: SF Has Enough Street Parking Spaces To Fill CA’s Coastline

Clarification: California’s coastline (840 miles) is shorter than the end-to-end length of SF’s on-street parking spaces alone (900 miles). This post originally compared it to the length of SF’s total public parking supply (1,451 miles long), which is actually longer than the United States’ west coast from Mexico to Canada (1,360 miles). Here’s a point […]

SFMTA Abandons Parking Meter Plans in Dogpatch and Potrero Hill

The SFMTA has given up on its neighborhood-scale plans to install parking meters in the Dogpatch and Portrero Hill, while parking meter approvals in the northeast Mission move forward at a snail’s pace. After two years of tangling with the city, the defenders of dysfunctional free parking have effectively caused a huge setback for progressive […]