Xerox- Reducing Congestion in Los Angeles through predictive modelling. Can data driven variable pricing really fix congestion in LA?

Xerox one of the largest equipment and services companies in the world is helping the city ofLos Angeles ease congestion through intelligent parking price management. Can data driven price changes fix congestion problems in LA? Or will consumers find a way to game the system?

Xerox – Reducing congestion in Los Angeles through Intelligent Parking Management.

Xerox company summary

Xerox is a US based global corporation that sells business services and document technology to companies and governments worldwide but it is most famous for its leadership in the photocopier business.  The company is less well known for its pioneering inventions of the computer mouse, desk top computing and other inventions that have since been further developed by Apple, Microsoft and other industry giants.

The problem in Los Angeles: Too much congestion

Motorists in Los Angeles were spending 10 minutes or more driving around, looking for places to park. The potential root cause of the problem? On-street parking prices did not match demand. Prices were uniform within a given area and were often the same, or cheaper, than garages a few blocks over. There was no incentive to park a little further away.

The project goal: change driver behavior and balance demand by achieving 10-30% of the parking spaces on each block being available throughout the day. Ensuring availability was expected to  reduce congestion and pollution, shorten travel times, and encourage the use of alternative forms of transportation.

The key question: Could analytics driven pricing be used as a mechanism to manage parking demand and ease congestion?

To answer the key question, the City of Los Angeles in partnership with Xerox commissioned LA Express Park. This one-year pilot program, funded by a $15 million grant from the U.S. Department of Transportation and $3.5 million in city funds, centered around a 4.5 square miles area  in bustling downtown Los Angeles.

To achieve the desired result, Xerox developed a highly integrated advanced pricing engine. This system  continually compiled occupancy and parking payment data, analyzed it, and pushed pricing data  out to customers (who were looking for available spaces) and enforcement officials (who were tasked with arresting and charging illegal parkers) in real-time.

The key challenges in developing the model

The key task was to design an appropriate model of how people choose parking spaces. This model would seek to optimize prices based on a history of observations. Development of  this model would be faced with numerous challenges some of which are discussed below:

  • It isdifficult to account for the immense flexibility that drivers have about where, when, for how long they stay and for what portion of their stay they are parked legally.
  • Populations react gradually, rather than instantly to price changes.
  • Drivers sometimes have to revise their decisions because the block they wished to park on is currently fully occupied.
  • Finally the whole system has a complicated dependence on its own history.

The Results

Despite the challenges, as early as the first six months of the trial, the LA Park Express pilot already realized substantial results:

  • Parking congestion decreased by 10%, with spaces easier to find
  • Under-utilized parking spaces decreased 5%, as drivers were drawn by lower rates
  • Pilot-wide parking rates decreased by 11%, but parking revenue increased by 2%
  • 76% of drivers indicated in a survey that they would park in less expensive areas nearby, demonstrating the power of pricing

Future sustainability

Are these early wins real? Can simply managing demand through changing prices fix congestion problems in Los Angeles? How can customers (parkers?) game the system? Is this method of congestion management sustainable?


Fitbit – leveraging consumers’ obsession with data


Gmail: ensuring a spam-free inbox with Machine Learning

Student comments on Xerox- Reducing Congestion in Los Angeles through predictive modelling. Can data driven variable pricing really fix congestion in LA?

  1. Interesting! I love the idea of being able to better track available parking spots – it seems like such an outdated challenge we face every time we park. However, I wonder how scalable this approach is. How does the technology in the parking spot actually work? Does it track spaces via satellite? Or through the meter? Does the 2% increase in parking revenue make up for the cost of installing this technology? I also wonder how much the prices have to vary to incentivize someone to park further away. For instance, if I could pay 25 cents more to park one block closer, I probably would, but for $1 more, maybe I wouldn’t? Interesting consumer behavior implications that I’m sure they could test as well.

  2. I wonder how the businesses reacted to the pilot. I could imagine that small business owners in popular areas, which receive benefits from foot-traffic in these popular areas, might have serious issues with this system. Imagine the local convenience store which suddenly finds that the parking near their stores is considerably more expensive that the rival 5 blocks away. Did they see a down-turn in business? Or did they find positive benefits: loyal customers were now able to find parking when before it was an issue. I would love to see how the broader parties (beyond the congestion obsessed city government) found the pilot

  3. This pilot is super interesting! I wonder if it’s scalable to other large cities where many people drive and have a hard time to finding parking (San Francisco, Chicago, Atlanta, etc). I like using price as a way to manage traffic and parking but I wonder how the parking companies feel. I’m sure they just want to have all spaces full at all times of the day.

  4. I grew up in LA and given the crazy amount of traffic there, I’ll glad to hear they are investing in this pilot! The results seem promising that using this price data an drive change in consumer behavior. I was initially skeptical that raising the prices would actual cause consumers to behave differently because parking is generally very cheap in LA, so I’m not sure how much more expensive they could go above $3 before people finding it unfair. I think the real benefit here would be if parking becomes more expensive, people might choose to rely on public transportation which would be a good thing for both congestion and the environment. Excited to see what comes of this pilot!

Leave a comment