Targeting 85 Percent Curb Occupancy Leaves One Space Open per Block: How Performance Parking Rates Eliminate Cruising Congestion

In early 2018, San Francisco expanded a pricing algorithm to all 28,000 of its parking meters, ending a decades-old system of static curb rates. The city abandoned the idea of charging a flat hourly fee across entire neighborhoods, shifting instead to a dynamic model that treats the curb as a managed commodity.[4][5]

The system, known as SFpark, relies on a simple mathematical target rather than a revenue goal. The algorithm adjusts the price of parking on every block to maintain an average occupancy rate of exactly 85 percent. If a block fills up, the price rises; if it empties out, the price falls.[3][4]

This demand-responsive approach solves one of the most persistent and invisible causes of urban traffic congestion. When curb parking is underpriced, spaces remain completely full, forcing arriving drivers to circle the block repeatedly. This phenomenon, known as cruising, clogs travel lanes and generates massive amounts of localized air pollution.[1][2]

Urban planners now recognize that static parking rates create a fundamental mismatch between supply and demand. "The price is too high if many spaces are vacant and too low if no spaces are vacant," wrote Donald Shoup, the UCLA urban planning professor who pioneered the concept.[1]

The High Cost of Cruising

Shoup formalized the 85 percent target in his 2005 book, The High Cost of Free Parking. He argued that free or cheap curb parking gives a temporary benefit to a lucky few while imposing severe social and environmental costs on everyone else.[1][2]

The core problem is that the supply of curb spaces is entirely fixed, while demand fluctuates wildly throughout the day. When a city charges the same price at 8:00 a.m. as it does at 7:00 p.m., it effectively guarantees that parking will be either severely overcrowded or entirely empty.[3][4]

Cruising for parking is not a minor inconvenience; it is a primary driver of gridlock. Studies conducted across multiple cities have shown that up to 30 percent of the cars in congested downtown traffic are simply circling in search of an open curb space.[2]

This small search time per car creates a surprisingly large amount of traffic congestion because, typically, many cars are searching for parking simultaneously during peak driving times. The collective fuel waste and carbon emissions from this daily ritual are staggering.[1]

By treating parking as a public utility that must be priced according to its fair market value, cities can eliminate this artificial congestion. The goal is to set the lowest possible price that still prevents a parking shortage, ensuring that the curb remains a revolving door rather than a long-term storage facility.[2][4]

The Goldilocks Principle of Parking

The 85 percent occupancy target is often referred to as the Goldilocks principle of parking prices. On a typical city block with eight curb spaces, an 85 percent occupancy rate corresponds mathematically to one or two vacant spaces at any given time.[1]

This specific ratio is the exact threshold required to eliminate cruising. If one or two spaces remain open on every block, a driver arriving at their destination can park immediately. The time spent circling drops to zero, instantly removing a massive volume of unnecessary traffic from the adjacent travel lanes.[1][3]

Achieving this balance requires cities to abandon the politics of parking and rely entirely on street-level data. Under a performance parking model, city councils no longer vote on meter rates. Instead, they set the 85 percent occupancy policy and allow the algorithm to find the market-clearing price.[1][4]

The adjustments must be frequent enough to respond to changing demand, but gradual enough to avoid confusing drivers. In San Francisco, the transportation agency evaluates occupancy data periodically and nudges prices up or down by small increments, typically no more than 50 cents an hour.[1][5]

This trial-and-error process eventually discovers the correct price for every block at different times of the day. A block located directly outside a popular restaurant might charge $5.00 an hour on a Friday night, while a block just two streets away might charge only 50 cents.[3][5]

San Francisco's Citywide Experiment

When San Francisco first launched the SFpark pilot in 2011, it installed wireless sensors in 8,200 metered spaces across seven neighborhoods. The sensors provided real-time occupancy data, allowing the city to track exactly how drivers responded to the shifting price signals.[3][5]

The results defied the expectations of critics who assumed the program was simply a municipal revenue grab. During the first two years of the pilot, the average hourly parking rate actually dropped by 4 percent. Prices declined in the morning and increased during the midday and afternoon peaks.[3][5]

Because the algorithm lowered prices on underutilized blocks, drivers began parking slightly further away to save money. This natural redistribution squeezed more blocks into the target occupancy range. The city reported that average cruising times dropped by 50 percent in the pilot zones.[3][5]

"What we've managed to do with SFpark is take parking management to the next level," said Ed Reiskin, the city's transportation director, when the pilot concluded. The success of the initial rollout gave the city the data it needed to expand the demand-responsive model to every meter in San Francisco.[5]

Other municipalities have since followed suit. Oakland implemented a similar demand-responsive system in its downtown and Chinatown districts, while cities from Seattle to Madrid have adopted dynamic pricing models to manage their own curb space.[4][6]

Reinvesting the Revenue

The transition to performance parking often faces intense initial resistance from residents and merchants who fear that higher peak rates will drive away customers. To overcome this, urban planners emphasize that guaranteed availability actually benefits local businesses.[1][2]

When curb spaces turn over more frequently, a single parking spot can deliver multiple groups of customers to a commercial district rather than being occupied all day by a single employee. The dynamic pricing model automatically lowers rates during economic downturns, keeping customers coming when demand drops.[1]

The most crucial element of the performance parking framework is how the generated revenue is spent. Shoup strongly advocated for returning the meter revenue directly to the metered district to pay for visible public services, rather than absorbing it into a city's general fund.[1][2]

When neighborhoods see their parking fees funding cleaner sidewalks, better street lighting, and improved public transit, the political opposition to paid parking typically evaporates. The revenue transforms the parking meter from a frustrating tax into a dedicated funding stream for local improvements.[1][2]

As more municipalities adopt the 85 percent rule, the next frontier is applying demand-responsive pricing to residential permit zones. Adjusting annual permit fees to balance neighborhood supply and demand could eventually guarantee that residents always find a space on their own block when they drive home.[1][6]