Welcome back to What’s NEXT, a podcast from Samsung NEXT exploring the future of technology. In this episode, I talk with Swiftly co-founder and CEO Jonny Simkin about using big data to improve urban mobility and how his company helps cities operate their public transit networks more efficiently.
Ryan Lawler: Jonny, welcome to the show.
Jonny Simkin: Thank you for having me.
Ryan Lawler: To start, tell us a little bit about what Swiftly is, and what you do.
Jonny Simkin: Yeah. We are a big data platform for transportation, primarily focused on public transit today, and what we realized is, public transit’s essential to moving people in cities efficiently, but a lot of the software and tools that they have available are decades old, and transit agencies do a lot of their analysis using Excel spreadsheets. What that leads to is, if you’re in a major city, you can load about one day of data or less, and then Excel crashes. What we’ve done is we built a new system. Completely modern infrastructure that takes very large data sets from cities, and we analyze them to find areas where you can improve operational efficiency, improve the schedule efficiency, looking at ways that we can find out which corridors or intersections are causing issues in the cities, so you can fix them. Really with the ultimate goal of improving the passenger experience, and also getting people from A to B faster.
Ryan Lawler: Okay, tell me about the founding story, like why are you interested in pursuing making public transit more efficient?
Jonny Simkin: My story is definitely different from my co-founder’s, but I never rode public transit growing up. I was born and raised in San Diego, Southern California. I drove everywhere, and I was always extremely frustrated with traffic, and congestion, and the question that I always ask is why you don’t just add another lane to the freeway, or to the road, because then you can have more cars, and the system should work better. What I realized when I was doing research on the subject is adding more lines typically just induces more traffic, and it’s not a sustainable way to solve the problem.
In 2012, I sold my first company, I moved out to San Francisco, and that was the first time I realized that owning a car was more of a burden than a sense of freedom. I sold my car, I noticed that most of my friends were selling their car, and I started relying entirely on public transit, and all of the new forms of mobility that you see around you. I fell in love with that way of life, and realized that public transit could be much more efficient, and devoted the last four years and hopefully several more years to come, to now seeing that vision and making it a reality.
Ryan Lawler: Okay. How did you even begin to understand this problem that cities face in making their public transit more efficient?
Jonny Simkin: It all started with my personal experiences. When I was taking transit there’s the schedule, and at first I was relying on the schedule because Google Maps only had the schedule data. It would say, “Hey, be at the stop at noon,” or whatever time I was supposed to be there. I was at the stop at that time, and then the bus or train just wasn’t there. I was like, “Okay, this is not a good experience.” Then I started using some realtime information from the electronic signs and in some of the apps that were being displayed in the city, and it would say, “Oh, based on realtime information, be at the stop in five minutes.” I’d be at the stop in five minutes, and that was wrong, and it was really frustrating as a passenger.
Our first instinct on solving the problem was really from the passenger side, which was just make transit easier to use. We worked on building a new mobile app and a prediction engine that took all the historical location data from buses and trains, and used that historical data to predict in realtime when the next bus or train would arrive, and to do so more accurately than existing systems. That was really impactful. In just a couple months we got about 10% of the city of San Francisco using our mobile app, but what we realized with that approach is even though it was really helpful for us as transit riders, it wasn’t necessarily a business in and of itself, and it was really tough to scale.
When we were talking with a lot of cities, what we realized was is even though each city is very unique, they all had very similar problems around making the schedule more accurate. Which is really important because that dictates their budget. How many drivers they need, how many vehicles they need, making their streets better support transit. We decided to build a lot of internal tools using all the data that we’re collecting so that we won’t just tell passengers, “Sorry, it’s running five minutes late.” Why don’t we work directly with the city to prevent that late issue from occurring in the first place?
Ryan Lawler: Okay, so you’re actually working with the transit agencies. How many cities are you working in?
Jonny Simkin: We currently have contracted almost 50 transit agencies primarily in the US. We work with agencies of all sizes, so our largest is Chicago Transit Authority. We work with Boston MBTA, Miami-Dade, we’re doing a very large deployment right now in Baltimore, and our smallest agency has two vehicles. Two buses all the way to over 2000 vehicles. The goal is really to make the system agnostic to the size of the agency to the geographic location of the city. We want it to be something that can be deployed anywhere in the world.
Ryan Lawler: What’s the hangup there for cities or transit agencies? What are the problems that they’re facing in being able to recognize that these lines are running late on a consistent basis, and to prevent that or make routes more efficient?
Jonny Simkin: I would say the most common problem statement we hear is, “We already produce a tremendous amount of data, we just don’t know how to use it efficiently.” Cities often times have systems in place. They typically have GPS or other form of tracking units in their vehicles. But what they do with all this data, when you start looking at it over days, or weeks, or months, or years, tends to be pretty limited because they don’t have the tools to really dig into the data. That’s really where we found the biggest problem is, is how do we take literally hundreds of millions of GPS coordinates gathered potentially per week in a city, and turn a raw GPS coordinate, which is, latitude and longitude’s not gonna tell you anything to focus on this intersection because one’s causing the problem.
Ryan Lawler: Take us back. How were these transit agencies operating beforehand? How would they plot out the routes from one place to another, or determine that a certain line should run down one street versus another?
Jonny Simkin: Almost all the major processes have been very manual in the industry. When it comes to route planning, they’ll take static census data, demographic data, and try to plot out routes that seem like they will be a good match for the general population characteristics in a given area or city. Once you implemented a transit network, it’s really tough to measure how well is it working. One of the things that you probably hear all the time with a lot of the tech companies that you work with, is iteration time. Really quick testing, you have a hypothesis, you test it, you roll out a new feature, or change the color of a button, and within in a couple of days, or a couple hours, you have a sense as to whether or not that worked, and you either build off that, or revert it.
That iteration process doesn’t really exist in transit, because of how long it takes. For example, you implement a new route, you said, “Let’s say it takes 60 minutes to get from point A to point B, so we’ll schedule based on that. 60 minutes we’re gonna add ten vehicles. The vehicles are gonna arrive approximately every six minutes.” But if it doesn’t actually take 60 minutes, then the whole characteristics of that route have changed. The whole cost structure, the number of vehicles you need have all changed. What if it only takes 50 minutes? Then you can potentially shave the whole vehicle and still provide the same level of service.
What agencies do today often is, they will send someone with a stopwatch to ride the bus and ride that same route a couple times and say, “Okay, well, we rode it five times, we think it’s actually is closer to 55 minutes.” But actually if you look at months and thousands of trips, that assumption that you got from riding the bus five times may not be correct. That’s really where Swiftly comes in, is looking at very large data sets to figure out how we should schedule more efficiently, or change routes.
Ryan Lawler: We’ve been talking very big picture, but let’s actually talk about the business. Tell me about what it’s like for you to engage with customers, what those conversations typically look like with a first meeting, and how do you sell these transit agencies that they need your product?
Jonny Simkin: The sales side of this whole business, and just selling to government in general is something a lot of people warned me about, and then I was really nervous about in the beginning. There are some challenges obviously with selling to government. You’re dealing with long and challenging procurement cycles. But, I found it to be in some ways extremely refreshing.
In these first meetings with transit agencies, it’s really clear that the people genuinely care about their city, and their residence, and you don’t see that a lot in the private sector. Mostly people are just focused on how am I gonna make that next buck. In the public sector, they’re often really focused on how can I do what’s best for my city? When we sit down, and typically the way that we sell the product is we go in with their actual data in our platform, and we say, “Let’s sit down, and tell us what are your current priorities and how can we just work with you on those priorities over the course of the next hour? Or, the next two days if it’s onsite workshop, and go over your actual data in our platform and see what that might be like.”
It’s through that, that there’s this aha moment where they’re like, “Wow, we’ve gotta work together because this is gonna be really good for our city, and we can do more changes faster, and we can iterate faster.” That was actually, believe it or not, a direct quote from when we were doing an onsite training, and they said, “You’re done more in the last 51 minutes than we have in the last two years.” That’s the typical way we sell, which is focusing on enabling them to be more effective, and I think that if we focus on making our customer successful, we’ll be successful too.
Ryan Lawler: Let’s talk about the types of data that you’re collecting. You mentioned GPS and realtime data for where different buses or other transit operations are at any given time. What else are you looking at?
Jonny Simkin: There are actually only two inputs to our system. The first is the schedule. Where are stops located, route paths, at what times are vehicles supposed to be there? Then we take the live GPS data, which typically is every 10 seconds to a minute. We then take that raw data and we calculate all of these new data attributes on top of that. Understanding vehicle speeds, understanding on-time performance. How is the vehicle doing relative to the schedule at this moment in time, and is that a trend we’ve observed in the past, or is that something that’s brand new? If you’ve ever been waiting for the bus and it says, “Oh, it arrives typically every 10 minutes,” but then it takes 20 minutes for it to arrive. That’s probably because two vehicles arrived at almost the same time, and then there’s 20-minute gap.
We can do all of these types of analytics just based on GPS. One of the things that we’re now hearing from our cities is, “You’ve shown me a whole new way of looking at GPS data that can boil down to improving my operations, improving my efficiency, improving the passenger experience. What if we gave you all these new data sets. Can you do something similar with them?” Now we’re looking at new data sets around understanding in large part, the movement of people in a city. How many people are getting on and off vehicles at certain stops? What time of day and how do we try to build networks that better map supply and demand?”
Ryan Lawler: Okay, and that is from data of people getting on to the bus and paying? How are you getting that information?
Jonny Simkin: A lot of different ways. The two most common are fare data. When you pay a fare, there’s a record of that. The other is what are called an automatic passenger counter. Most fleets, at least a percentage of them have APCs, which are laser beams that are tracking on and off counts. You can have a sense as to how many people got on or off at a given stop, and how many people are in the vehicle at any given moment in time.
Ryan Lawler: One of the great quotes I think in startup land I think I saw at first with Alex Rampell at Andreessen Horowitz. Says, “The battle between every startup and incumbent comes down to whether the startup gets distribution before the incumbent gets innovation.” I think this is so true in urban mobility and urban transit because over the last five to 10 years, we’ve seen a real change with the ride hailing apps and everything else is showing up in the cities, so I’m curious how you think about that. If you have a bunch of private enterprises, that are just out innovating or moving so much faster than public agencies are able to, what does that mean for the future of public transit?
Jonny Simkin: A very interesting question. Transit agencies are dealing with arguably one of the most evolving transportation landscapes in the history of mankind. You have autonomous vehicles, you have all of these new modes whether it’s ride hailing, or bike share, or car share, or now scooter shares all over San Francisco in many of the major cities. How does public transit fit in to this? I’m a firm believer that public transit has a very strong place in both the mobility options of today, and in the future. I think if you look at just general population sizes, you need mass transit to be successful, and to be able to move people without congestion. But, undeniably public transit has to adapt. I think a large part of that has to do with making decisions based on data not gut. Making policies also based on data and not gut. I think most of the cities we talked to are moving in that direction very quickly, and it’s really encouraging to see that.
But, the other thing I would say is, even though private companies … and you think Uber or all the car share providers, or scooter share, bike share … they’re evolving really quickly, but they offer a very different product than public transit. I would argue you often don’t want public transit to adapt so quickly that it’s no longer safe and reliable and affordable. There’s this right balance that I think public transit needs. I would say the industry itself is trying to find that balance, but I do see a really strong future for public transit because let’s look at the numbers for example. Five billion people in urban areas by 2030. You can have 100 million Ubers on the streets, but if you have one person per car, the numbers don’t make sense and if you just look at the width of our streets, the infrastructure is also limited. You can’t just change street widths.
I think where public transit has a really unique place, is getting to move very large volumes of people very quickly and efficiently, and a lot of that comes down to having access to streets in a way that private mobility option don’t have access.
Ryan Lawler: We’ve been talking a lot about these short-term direct effects of using data to make transit more effective and more efficient, but can you talk a little bit more about some of the second, or third order effects that come from better transit within cities.
Jonny Simkin: The current estimate is two to three percent of GDP is lost by wasted time in congestion. I’m not gonna claim we’re saving the US two to three percent of GDP, but, when you think about that way that, “Holy crap, hundreds of billions of dollars are lost but people just sitting in the car doing nothing. They’re not at work, they’re not producing whatever they do at work, they’re not spending money with their friends, they’re just in the car doing nothing.” If we can make transit more effective, and we can reduce congestion, that’s a massive impact on just the economic productivity of the country, or any country that we deploy our services in.
There’s also the environmental impact. Cars produce about a fifth of greenhouse gasses in the US. There’s a huge impact where we can get people out of their cars and taking mass transit, we can improve the environment. And then, there’s the quality of life which isn’t necessarily as quantitative, but I can tell you qualitatively speaking, having moved to San Francisco, selling my car, and being able to walk and take transit everywhere, I’m much more happier person, and I don’t complain or swear when I’m in traffic anymore. I think three huge longer term benefit that are important to keep in mind.
Ryan Lawler: What’s one controversial opinion that you have that’s really strongly held?
Jonny Simkin: I don’t know if this is controversial, but this is something I’ve been thinking a lot about recently. It’s very cliché, but I think it was probably middle school. I heard the Gandhi quote, “You must be the change you wish to see in the world,” and that’s something that I’ve been spending a lot of time trying to be. What I’ve realized recently is that I now think about it as, you must enable others to be the change you wish to see in the world. That is if I am focusing my time and my effort trying to accomplish my agenda, I’m not nearly as powerful as if I empower others to do that. As I’ve been hiring people to join my team, I’ve realized if we’re going to be successful as a team and the vision’s going to be successful, it can’t just be me. It just can’t be one other person, it has to be a team of people. Now I spend a lot of my time thinking about how can I enable others to be successful.
Ryan Lawler: Let’s pull that thread a little bit, because I think that in startup world, there’s this credo of asking for forgiveness, not permission. But, by the nature of the types of agencies you work with, you can’t really operate that way. Right?
Jonny Simkin: Interestingly enough, I was in Austin two days ago meeting with some customers there. What was really interesting is that what you just described, I think is changing. I was talking with one of the executives at one of the transit agencies we work with, and he said, “Up until now, the agency’s been risk averse, but we’re not going to change anything, because if we do, there’s a risk of getting in trouble. If we don’t, we for sure won’t get in trouble.” He said, “In order for us to be successful moving forward, we have to take those risks, and in order for those risks to be successful, they have to be based on data, or some real world information.” Then we started going down that path, and then last night, another one of our top 15 city in the US, I was meeting with the assistant director there, and he said the exact same thing that I had just heard in Austin two days ago. A lot of people do think that way, and I think it’s totally changing now and it’s very exciting.
Ryan Lawler: It’s changing from the public agency perspective.
Jonny Simkin: From the public agency perspective, yes. From the management in the public agency, and now it’s starting to permeate to their staff as well.
Ryan Lawler: Okay. If you weren’t doing this, what areas of technology would you find most interesting, or do you think that you’d like to work on?
Jonny Simkin: I actually love the transportation space. I can see myself staying in it for a long time whether that’s through AVs, or obviously the impact of AVs on public transit. I think there’s just so much change that it’s really interesting, and it touches everyone. Healthcare, not glamorous, but really high impact. Energy, again not glamorous but really high impact. I tend to gravitate to the less glamorous, high impact, types of industries.
Ryan Lawler: When you talk about AVs and autonomous vehicles, I find it really interesting what impact they’re going to have over the way that we get around. Not just in cities, but really anywhere. I’m wondering if you have a thesis or an idea on how they’re going to change the way that we think about transportation, the way that we think about car ownership, the way that we think about even where people live.
Jonny Simkin: I think AVs will change society a tremendous amount, and that’s an understatement. You touch on car ownership, some people will probably want to own a car but, it really doesn’t make sense when that asset is idle 95% of the times, and I don’t see myself owning a car again. I guess it depends, but probably I won’t own a car again. I do think the one thing that AVs can’t necessarily touch, is density of restaurants and things around you. AVs might make it more accessible to live out of the city, you can work while you’re driving, but at the end of the day, if you want to just get out of your house or apartment, and walk to a hundred bars, or a hundred restaurants, or a hundred things to do in the city, you need that density. I do think AVs won’t necessarily adjust that transportation as the fabric to the people and places around you, and if transportation is going through a huge evolution, so will everything else.
Ryan Lawler: I like to think of also how it will change our cities. I think about San Francisco in particular and I know that the city has changed a lot even since the earthquake when they took down the highway along the highway along the Embarcadero and through Hayes Valley or whatever, and I think, “How did would the city be if we didn’t have that overpass running down 13th street, or Duboce, or whatever it is?” How much of the city could be reclaimed if we had smart vehicles that didn’t have to bunch up onto this freeway?
Jonny Simkin: I think about that all the time. Or, what about parking lots? Parking structures? How much real estate is being wasted by cars just sitting in there doing nothing? The boom that it’s gonna have on retail and real estate values is gonna be tremendous because now that parking lot could be an apartment building and maybe rent will go down a little bit in the city. Or, at least you’ll have another restaurant, or bar, or thing to do around you. You think about transportation again, being this fabric of the people and places around you, and as that changes everything else will too. The places around us, will definitely change.
Ryan Lawler: Talk me through the big vision. What happens when every city is using Swiftly? How does it change the way we get around? How does it change what those cities look like?
Jonny Simkin: Our mission is really to make cities move. Just three words, make cities move. That has efficiency tied into it, reliability tied into it, time savings, productivity increases. I think if Swiftly were to deploy in a really deep way, you’ll start to see transportation networks evolve to meet the needs of the residents and the visitors. Not the other way around where residents and visitors have to try and follow what has already set in stone.
Ryan Lawler: All right. This was great, Jonny. Thanks again for joining us and telling us more about Swiftly.
Jonny Simkin: Thank you so much for having me. This was great.
Originally published at https://samsungnext.com on June 28, 2018.