Why ‘On Time’ Isn’t Always on Target

00:00 Stop Requested.

00:01 Levi McCollum: Welcome to Stop Requested, the podcast where we discuss everything transit. I’m your co host Levi McCollum, director of operations at ETA Transit.

00:10 Christian Londono: And I’m your co host Christian Londono, Senior Customer Success Manager at ETA Transit.

00:20 Levi McCollum: Hey Christian, how are you today?

00:23 Christian Londono: Doing good, Levi, how are you?

00:25 Levi McCollum: Hey, I’m doing pretty well. I’m really excited about this one. I know we say that each week, but on time performance is this week’s topic. And the reason why we’re choosing on time performance for our listeners is because last week we talked about unlinked passenger trips or ridership and we discussed how important a metric that is in the transit world. We’re going to shift gears and talk about another metric that we think is also pretty important in the transit space. So just to kick off the conversation, Christian, what do you think of on time performance? How do you define it?

01:04 Christian Londono: Yes, and again, you know, we talked about last week whether ridership was the most important metric for public transportation—and you know, maybe it is, maybe it’s not. It’s hard to pinpoint just one metric to measure public transit. But if ridership is first, I would say on time performance would be second, simply because of the impact it has, especially for the public. To get started and answer your question, I’ll give you a quick definition for our listeners: on time performance refers to the percentage of transit vehicles—buses, trains, any transit vehicle—that arrive at their scheduled stops within a specific window. It could be a few minutes or many minutes, as defined by the agency and the mode of service, and it is crucial for reliability. It really helps maintain trust in public transportation. In essence, it’s almost like a contract with the public: a transit agency is stating that its service is going to arrive at a certain time, and that becomes the expectation for riders. The calculation is generally done by dividing the number of on time arrivals by the total number of scheduled arrivals, often using a window of, say, one to five minutes for fixed-route service. I think it’s very important for our riders.

03:02 Levi McCollum: Yeah, that’s interesting that you honed in on the arrival portion rather than departures. Can you explain why you think arrivals should be considered part of that measurement rather than departures?

03:15 Christian Londono: That was kind of the overall definition of how it’s calculated. But, you know, thank you—going there because usually transit agencies (and I’m mostly referring to bus service) capture both arrivals and departures. Both are very important. However, for this calculation, agencies often refer more to the departure piece because after the arrival–departure event takes place, on time performance is sorted into three buckets: early, on time, and late. The departure piece becomes very important because arriving early at a stop might seem good—after all, it’s not late—but if the bus departs early, passengers may miss it. For example, if a bus scheduled for 10:00 am leaves at 9:58 am, a person waiting for a 10:00 am bus could miss it even though the bus was technically early. So thanks for bringing that up. With that in mind, what’s your experience regarding the calculation and the use of those two timestamps—the early and the late, the arrival and the departure?

05:30 Levi McCollum: In my experience at Fixed Route Motorbus, the calculation has been based on departures. I’ve heard from our customers at ETA and colleagues around the nation that departures are the key metric. I think you’re absolutely right: if a vehicle arrives early, that’s acceptable as long as it doesn’t depart early, because a passenger can still board the vehicle even if it arrives two minutes early and then waits until the scheduled departure. But if a bus both arrives and departs early, that’s a serious issue. I even heard from agencies—just the other day in Greenville, North Carolina—that there’s no excuse for departing a time point early. This leads me to another part of the question: what would you consider to be the on time window? In my experience, it’s about six minutes—60 seconds or less would be considered early, then you have from 59 seconds before the scheduled departure up to five minutes and 59 seconds after the scheduled departure as “on time.” Anything beyond that is late. Is that your experience as well?

07:07 Christian Londono: Yes, that’s typically the benchmark in terms of the calculation. What’s interesting is that some agencies claim a 0-to-5‑minute window because they give themselves 59 seconds on the early side; when it hits 60 seconds, it automatically counts as early. So if it’s 59 seconds ahead of schedule, it’s on time, but if it reaches one full minute ahead, it’s automatically considered early. That’s the “early” bucket. Similarly, if a bus is up to 4 minutes and 59 seconds late, it’s still on time, but beyond that—say, a 10:00 am bus arriving at 10:05 am—it becomes late. Some systems even use decimal points (like 1.2 minutes or 4.8 minutes) to allow for a small wiggle room. There are variations: some say 0 to 5 minutes, others might effectively be 1 to 6 because of that extra 59 seconds. In practice, when you dive into the data, some agencies might give themselves a bit more room in the early or late buckets. And to add another point from what you mentioned about Greenville—there’s often no excuse to be early, yet drivers sometimes can’t do anything about it. In some cases, if they try to avoid being early, they might end up departing late at the start of their trip because, in certain locations, they cannot safely stop for the required time without disrupting traffic. Sometimes bus operators, when forced to wait (maybe even for a bathroom break), may get signaled by a police officer to keep moving. So schedules can be inherently unfair for bus operators, putting them in a position where they might either be forced to be early or, conversely, run late to compensate. On time performance is such an important metric because it reflects the quality of service as perceived by commuters. If a bus is consistently delayed, people get frustrated—they might miss appointments or arrive late to work. I even work in customer service for a transit agency, and I’ve heard riders say, “I’m going to lose my job because of this,” or “If I get fired, you’ll have to pay my bills,” because the service isn’t on time. That dissatisfaction is directly tied to on time performance, which in turn affects reliability. But being late can create a host of issues. Levi, what do you think about that?

11:53 Levi McCollum: Yeah, it certainly can cause frustration and even discourage people from using public transportation altogether. On time performance is critical because it directly impacts the rider experience. One thing you mentioned that I’d like to revisit is the importance of having a safe layover. Safety is paramount for transit agencies—you need to deliver people safely. If you’re unable to hold a safe layover without impeding traffic, you’re in a tough spot. You want to keep moving, but if you depart too early, you risk penalties for not being on time, which can affect your performance evaluation as an operator. Many agencies don’t measure on time performance on a stop-to-stop basis; instead, they focus on key time points along the route. In many cases, there might only be four or five such opportunities. These critical time points need to be chosen carefully to ensure there’s a safe layover and that the schedule can accommodate external variables like traffic. There’s so much to consider, and it really highlights the challenges of not just measuring, but actually achieving reliable on time performance.

14:09 Christian Londono: Yeah, achieving perfect on time performance isn’t easy. People might wonder, “Why can’t you just be on time all the time?” Poor on time performance, influenced by many factors, can lead to confrontations on the bus. I’ve seen situations where passengers board a bus that’s running late and then take their frustration out on the bus operator—blaming them entirely for the delay. But there are many obstacles to maintaining, say, a four-minute on time window, especially for bus services that aren’t BRT. Regular bus service in mixed traffic faces issues like heavy traffic, accidents, road closures, and weather (hurricanes in Florida or heavy rain, for example), all of which can significantly impact time performance. Levi, can you think of any other factors that affect on time performance?

16:06 Levi McCollum: I think you nailed the big ones. We probably have it relatively easy here in Florida compared to, say, Texas or Minnesota. I remember one Friday night on the DART system in Texas—there were heavy winds that broke a lot of branches, which fell on the rail line, forcing the agency to clear them; you can’t just plow through large branches. And up in the north, like Minnesota, dealing with a lot of snow and icy conditions makes it even harder to keep on schedule.

17:11 Christian Londono: Yeah, accidents also play a big role. In adverse weather like heavy rain or snow, there are more opportunities for accidents and collisions. Even if the bus isn’t directly involved in an accident, a blocked road can delay service for hours—or even cause a trip to be missed altogether. It’s important for transit professionals to understand that while riders often focus on getting where they need to go, many factors outside of the agency’s control—like mechanical issues—also affect on time performance. Agencies track these as “incidents,” and while some are within their control (like a mechanical breakdown that recurs because of poor maintenance), others are not. All of these factors can impact how reliably the service is delivered.

19:17 Levi McCollum: There’s also variation in how an operator drives the vehicle. I’ve been on buses where you really have to hold on because the driver is aggressively punching the accelerator with their left foot. In contrast, some drivers take a more cautious approach and accelerate slowly, prioritizing safety over speed. There’s no judgment there—it just shows that operator style can affect on time performance.

19:53 Christian Londono: That’s a big one. In my experience, most bus operators I’ve encountered take great pride in keeping their routes on time. Some drivers run a bit “hot” to finish their routes quickly, especially if they want extra time at the end of the line, or if they’re in a hurry due to personal needs. Conversely, some err on the side of caution and drive more slowly to ensure safety, even if it means arriving a few minutes late. Both approaches affect on time performance.

21:01 Christian Londono: I also want to touch on benchmarking. Most agencies use a 1-to-5 or 1-to-6 minute window (or sometimes 0 to 5 minutes) for fixed-route service, based on automated vehicle location (AVL) systems like GPS to compare actual times with the schedule. For example, if the schedule says 10:00 am, the AVL data will determine whether the bus arrived between 10:00 and 10:05, or if it arrived later, say 10:08 or 10:10, which would count as late. That’s the basic calculation. In terms of benchmarking, Levi, what do you think is a good on time performance for a fixed-route system? Should a system be 99% on time to be considered good, or is something like 60% acceptable? What’s your take based on your experience and data from various sources?

22:34 Levi McCollum: This is a hard one to answer because there are so many intertwined factors. First, you have to consider the mode—are we talking rail, bus in mixed traffic, BRT, or paratransit? Then you have to look at it route by route. Is it a headway-based route or schedule-based? Those factors differentiate the calculation. If I had to pin it down for fixed-route bus service in mixed traffic, my experience tells me that around 80% on time is what you should aim for. Of course, that window can vary—are we using a window of negative 59 seconds to 4 minutes and 59 seconds, or a 10-minute window? When an agency claims 85% on time, there are deeper questions to ask: What window are they using? Are they looking at arrivals or departures? It often prompts more questions than answers. I wish there were better standards for on time performance, but it’s very community based. What do you think, Christian? Do you have any more structured thoughts on this?

24:36 Christian Londono: I’m exactly where you are. There are different factors at play, so it’s not a cut-and-dried answer. Typically, for a fixed-route system (not BRT) operating in mixed traffic, you’re looking at a goal of around 80% on time, with a stretch goal of 85%. If an agency is at 85% or above, that’s pretty good, especially considering factors like mixed traffic, accidents, and the lack of dedicated lanes—all of which are outside the agency’s control. Aiming for 100% isn’t realistic. I’ve seen performance reports where a fixed-route system boasts 90% on time performance on paper, but when you dig deeper, there aren’t enough critical time points to match rider experience. So there are layers to this. But if you’re using that zero-to-five calculation with well-situated time points, I’d say 80% is a really good goal for fixed-route service. Then, of course, there are other modes to consider.

27:15 Levi McCollum: In my experience, demand response (paratransit) typically has a goal of around 90%. There are many reasons why this might be higher than traditional fixed-route bus service in mixed traffic. For example, both Li Tran and Palm Tran use 90% as their goal. As for rail, I’d imagine it’s at least 90% given that it operates on a fixed guideway with fewer interruptions—provided preventative maintenance is up to par. I’m not 100% sure on rail, but that’s my sense of it.

28:12 Christian Londono: Yeah, not a lot on rail—I agree. What’s interesting about rail is that if there’s an accident on the road affecting a fixed route, you might be able to deploy a detour and bypass the problem area, with appropriate rider communication. But with rail, there’s typically no detour; you just have to wait until everything is cleared, which can have a greater impact on service. Regarding demand response, especially in paratransit, riders are often given a 30‑minute window. For instance, if a rider books a trip and is told their vehicle will arrive between 10:00 and 10:30 am, then as long as the bus arrives anywhere within that window—even if it’s 10:00, 10:10, or 10:30—it’s considered on time. That 30‑minute window is much larger than for fixed-route service, which is why paratransit can target a 90% on time performance compared to the 80% typical for fixed routes.

30:37 Levi McCollum: That’s some good insight, and I know you have direct experience from working in paratransit. It’s helpful to hear that. What do you think are some future characteristics of on time performance? Are there ways it could improve thanks to technology?

31:05 Christian Londono: I do think so. Many transit agencies are starting to apply new technologies. For fixed-route on time performance—as well as for other modes, especially when operating in mixed traffic—the schedules and running times aren’t static; they need constant updating as community factors change (like a new Walmart or employment center affecting traffic patterns). Agencies are increasingly analyzing real‑time data and adjusting schedules accordingly. In terms of technology, many systems are working on transit signal priority (TSP), which helps vehicles get through intersections faster by keeping the signal green longer for buses. This can reduce travel times and improve the chance of arriving on time. Additionally, some agencies are exploring “light BRT”—not full-scale BRT with entirely dedicated lanes but segments of prioritization that can improve travel times and reliability at key points along the route.

33:33 Levi McCollum: I certainly think that both full BRT and light BRT can be useful tools. I would never argue against giving buses dedicated lanes in most cases. However, there’s a bit of a debate these days because some systems are classifying “light BRT” as if it were full, proper BRT—and that can water down the benefits. We have to be careful with our terminology. Ultimately, though, we’re all advocating for more dedicated transit lanes.

34:35 Christian Londono: It’s difficult, right? From the perspective of a transit agency or planner—and as someone who grew up in a city with a regular fixed-route system operating in mixed traffic that later transitioned to a robust, full‑scale BRT—the difference is transformative. When a proper BRT system is implemented, people start switching modes because it’s so much more convenient and faster than driving. Ideally, transit agencies and professionals would be given the opportunity to prioritize public transit more, but in the U.S. the car-centric culture and land development patterns make it tough. Often, agencies have to start with a watered‑down version of BRT, but you have to be careful because calling it “BRT” when it’s not can tarnish the concept if customers come to expect full on time performance from it.

36:34 Levi McCollum: The implication is that you don’t get the full on time performance benefits of a proper BRT system when you’re running a “BRT light” operation. If you call it BRT light, you risk tarnishing the reputation of BRT overall—especially if riders expect a fabulous, on‑time service and then experience delays. It could lead to conversations about why BRT isn’t performing as expected and why it can’t stick to its scheduled departure times. So we have to be cautious with our terminology.

37:28 Christian Londono: Yeah. And I think I wouldn’t even change that goal.

37:32 Levi McCollum: Right.

37:32 Christian Londono: For example, if you’re a fixed-route system and one or some segments start to get dedicated lanes or resemble BRT or light BRT, I’d keep those performance goals closer to 80% because it wouldn’t be fair to measure a watered‑down BRT the same way as a full‑scale BRT. Overall, though, this is an exciting conversation about service reliability and on time performance. As we stated at the beginning of this episode, it’s one of the most important metrics—especially from the customer perspective. Customers might not care about overall ridership numbers, but they certainly care about being on time. And on time performance can vary by route—some heavily traveled routes might have excellent performance while others may lag behind, which affects the overall rider experience.

39:34 Levi McCollum: It really is like peeling an onion. I’d love to see more standardization around on time performance, though I doubt it will happen since schedule adherence is such a local topic and varies by geography. If I were king, I’d push for a more holistic calculation to determine transit agency health. But for now, we’d love to hear from you, our listeners. This discussion will be posted on LinkedIn—let us know in the comments what your experience with on time performance is, and if you have any feedback or ideas for future topics, please share them. We’ll catch you next week, and thank you for listening.

40:42 Christian Londono: Thank you for listening. Thank you, Levi.