As bike share companies added more bikes to Seattle streets during late summer and early autumn, the number of trips accelerated, too. Combined, Spin, ofo and LimeBike are carrying nearly as many daily riders as both the First Hill Streetcar and the South Lake Union Streetcar combined.
For the first two months, the brand new bike share services carried just under 120,000 trips. Since then, as the number of bikes ballooned, the number of trips accelerated even as typical seasonal biking trends say ridership should have been slowing down. By the end of November, the most recent figures Seattle Bike Blog has received from SDOT, the combined trip total reached 347,300. Riders even passed the million-mile mark.
To put that in perspective, Pronto Cycle Share carried 278,143 trips total during its entire two and a half years of operations. It took private bike share companies just over three months to pass Pronto’s lifetime total. The number of bikes permitted was about 9,400 by the end of November, though many of those are not in service due to maintenance. Pronto had 500 bikes.
At this point, Pronto is no longer a useful measuring stick for bike share in Seattle. Averaging 2,711 rides per day, bike share companies have already blown past ridership on the First Hill Streetcar (1,600 riders per day) and the South Lake Union Streetcar (1,400 riders per day). In fact, considering the accelerating rate of ridership, it’s likely that the bike share companies have had many days carrying more riders than both streetcars combined.
That bike share services are already rivaling two streetcar lines that cost about $190 million to build is pretty incredible. Free-floating bike share services have cost the City of Seattle almost nothing. In fact, these companies pay the city permit fees. And at $1/ride, bike share is the cheapest way to get around other than walking or owning your own (not-high-end) bicycle. That’s far, far cheaper than other private mobility services, like car share, taxis and app taxis.
Healthy, zero-emission, popular and affordable transportation that could lower traffic congestion at nearly no cost to the city? I understand that this all sounds too good to be true. But so far, that’s what’s been happening with free-floating bike share in Seattle.
It’s not all perfect, of course. Bikes that block walkways (whether left there by users, blown over by the wind or pushed over by drunk people) cause problems for people with vision impairments and people who use mobility devices.
But with such amazing benefits coming from the bike share services, the city should be looking at how to help these services succeed and keep walkways clear. And some reasonable investments could go a long way to help. The good news is that SDOT will soon be trialing some designated bike parking areas specifically for free-floating bikes.
My favorite idea is for SDOT to expand its existing on-street bike corrals to include extra space for free-floating bikes. The city should also add a lot more bike corrals all over the city. On-street bike parking helps keep sidewalks clear, and even if a bike is knocked over, it likely won’t fall into the walkway. Bike corrals can also be installed near street corners, which helps prevent illegal car parking that often encroaches on crosswalks and blocks sight lines. That’s good for everyone.
Imagine if every intersection in a commercial or designated pedestrian zone had at least one bike corral. This would not only help keep walkways clear, it would also dramatically expand available bike parking for people’s private bikes.
As the pilot permit stage is set to wrap up December 31, the city does not yet have a permanent bike share permit ready to go. Instead, companies will continue operating as they have been while SDOT develops a permanent permit over the winter. Details from SDOT:
SDOT will evaluate data from past six months to determine if free floating bike share permitting is feasible in the City of Seattle. The Transit and Mobility team will evaluate:
- Ridership Data
- Safety and Collision stats
- Vendor Compliance
- Bike Parking issues
SDOT is concerned with improperly parked and misplaced bikes, and as part of the final evaluation, is actively working on a solution to correct the issue. Residents can expect a new parking initiative in the coming days that provides designated parking spaces for bike share bikes to address parking issues for during the remainder of the permit year. We expect to launch in three neighborhoods in the coming days…
- Total bikes: 9,388
- Average Trip Duration: ~30.3 minutes
- Total Trip Duration: 7,306 days
- Average Trip Length: ~3.0 miles
- Total Trip Length: 1,052,061 miles
- Total Trips: 347,003
- Average Trips per Day: 2,711
- Weekend Trips Percent: 33%
We should be getting updated data, including trips through December, at the end of the month. So stay tuned.
Comments
47 responses to “Bike share pilot’s daily ridership blows past Pronto’s lifetime totals, rivals both streetcars combined”
the streetcar ridership is a low bar.
The place where the current model falls below Pronto statistics is rides per bike. However I’m guessing that there is a pool of bikes that remain idle in out of the way spots for many days. It’ll be interesting to see where the usage hotspots are. Around UW the bikes are still moving quite a bit even in December from my anecdotal observations.
Is overall bike ridership up? I couldn’t see any support in the Fremont bike counter. However I don’t think that bikeshare is displacing that many rides that would have otherwise been taken by rider-owned bikes. Certainly my bikeshare rides have been in addition to my regular commute, but maybe for others the story is different.
Another thought regarding rides per bike per day: one reason it is a leading metric for stationed bike share systems is that the metric measures system efficiency, which is important for a lot reasons, but mainly because the great costs to the city per bike.
Is this important for a free floating system, where the bikes are free to the city? Definitely still important, especially if we worry about financial viability of the companies, but is it as important? I’d be interested to hear thoughts on how SDOT should measure success of it’s current system
It’s hard to find support for much in bike-counter numbers because weather is such a dominant factor in bike usage. I’ve done some quick-and-dirty analyses that suggest the Westlake Cycletrack has accounted for an increase of a few percent in cycling on the Fremont Bridge, but my method of neutralizing year-over-year weather differences (comparing Fremont’s performance vs. previous years to other counters’ performance vs. previous years) wouldn’t work for a change that affects the whole city, like the introduction of dockless bikeshare.
There was an STB post in the past that modeled bike counts based on publicly available weather data and a few other things. That sort of analysis is what we’d need, but we’d need to wait longer to be sure of anything.
Al, I just completed a project using the bike counter data, with my project partner Cesar (an ex-transportation planner). Among other things, we used a Facebook algorithm to decompose the data into daily, seasonal and overall (aka secular) trends.
Here’s a directory of pictures of the results :
https://github.com/cspitmit03/data602-finalproject/tree/master/web/trend
Here’s Fremont:
https://github.com/cspitmit03/data602-finalproject/blob/master/web/trend/Fremont.jpg
The results are adjusted for temperature and rainfall.
PS If anyone knows someone who can host the results on the web in a ‘live format’ let me know, the code just needs to be reran monthly when the new results are released. I expected but did not see much of an increase in the counter numbers. The other features of the project were a graphical display of historical figures with various filters and perspectives, as well as a weather-based ‘7 day bicycle forecast.’
@Asher: Cool stuff. What’s your source for all these other counters? In the data I found linked from City of Seattle websites I found that basically only Fremont, Spokane Street, and 2nd made any sense — the other counters either had long periods where they appeared to be out-of-order, or that there were year-to-year changes that didn’t track with weather as expected… indicating they were probably dominated by changes in counter calibration or something like that.
FWIW my super lazy analysis is here. I’m only using month-by-month counts. The (inaptly named) “Residuals” tab is the one where I compare each counter against past performance. 2nd Ave doesn’t have enough data yet. The really notable events it brings out, where the two bridges performed differently, were the Alaskan Way Viaduct closure (causing unusually high bike usage on the West Seattle Bridge, as it messed up cars and buses), and the Westlake Cycletrack (a prolonged period where Fremont did worse during construction, then better after completion).
Al,
I started the trend analysis at the date each counter began. For missing data, I identified the counter most highly correlated with the given counter, and then used a ratio to that counter to impute the missing data. E.g. if Second street was most similar to Fremont, with on average a quarter of the traffic, I’d multiply Fremont’s count for that time by one quarter and use that as the result for Second. For Fremont, there was very little missing – I used the average of observations a week before and after for that.
I’m still surprised the counters show no impact from dockless bikeshare, not just in the trend analysis but in the absolute totals.
I’ve never been to Seattle, so I have very little local knowledge.
The system has 18x the number of bikes that Pronto had with a much bigger geographical footprint. It shouldn’t be surprising to see these numbers. They alone cannot define success.
Dockless looks weirdly inefficient: 2711 trips per day divided by 9388 bikes? Less than .29 rides per bike per day. Pronto: 278,143 total rides divided by 29 months (approx) of operation = 9,591 rides/month. 9,591 rides divided by 500 bikes = 19.18 rides per bike per month, divided by 30 days in a month = .64 rides per bike per day, or twice the per bike ridership of dockless. Would I say that therefore Pronto was the better system? No. It lacked a large enough service area and it probably should have had an e-bike component from the start. But if your main argument is that the numbers are great, you need to put those numbers in context.
Re the streetcar as a low bar — it is more akin to Pronto with its limited footprint, so it’s not surprising that dockless will look good by comparison.
I see constant reports of people having to try 2 or 3 or more bikes before they find one that isn’t broken, or reserved by another user, or otherwise inoperable. I’m baffled why this isn’t included in your brief mention of the downsides. It’s also unclear that the business model on which $1 rides are predicated is sustainable, and whether these firms will survive – would that leave the city once again without a system? Are these companies selling data gathered via their apps? These are all serious potential downsides that you’ve omitted.
The 2711 trips per day number is a misleading one, since that’s averaging over all the days since program was launched in July. But for the first 2-3 months there were a much smaller number of bikes deployed than there are now.
Ron is right. These are total trips since the pilot began. There weren’t 9400 bikes when the trial started, more like ~1,000.
Earlier statistics put rides per bike per day at 2.56 [1].
https://www.seattlebikeblog.com/2017/09/20/bike-share-ridership-is-booming-spin-launches-better-bike-forbes-says-limebike-valued-at-200m/
Kimberly’s point on the sustainability of the business is a great one. 0.3 rides/day @ $1/ride is $100/year (and that assumes the 0.3 is the average even including winter months). Is that enough to pay for these bikes, including transaction fees, maintenance, vandalism, and theft, over their relatively short usable lifetimes before they need to be replaced?
Offhand, I’d guess they need at least three times that, which may be possible, but is a high hurdle, especially given the low usage you’d expect in winter months. I am rooting for these companies, but it’s important to be realistic about the challenges they face.
That doesn’t even get into the issue of how many of those rides are paid vs promotional codes. Pronto was pretty stingy on the public promo give-aways whereas the dockless bikeshares have several promo codes active right now. Ofo is free until Jan 2, Spin has SANTACON2017, BIKETOPIKE in just the past week. Lime probably has their own.
If you want to dig down into the financial model how many rides are paid will matter.
I think right now, the companies are focused mostly on growing the system, in hopes of getting that critical mass of users to turn profitable. I don’t think their goal is to be profitable immediately from day 1. But at least they do seem to realize that bike share (and vehicle sharing in general) is a go-big-or-go-home business. You just can’t get to that critical mass with just 500 bikes and an area the size of Pronto’s.
With all due respect to Pronto and its defenders….
Floating bike share has huge advantages, including the fact that you can use them, even if you live/work outside the very small urban core where Pronto had stations.
Perhaps an ideal solution would be both docked and floating bikeshare. But if a city has a choice between the two, floating bikeshare is the way to go.
The last report on Sept 19 reported 2231 rides per day and 118,240 total, so the analysis was for the first 60 days. The current analysis reports the first 128 days. So that means that there have been an average 3364 rides per day over the 68 days since the first report.
Since Sept 19 UW went back into session – and UW students use these bikes a lot – but the weather certainly got worse. The bike numbers also doubled.
I wouldn’t expect the Freemont bike counter to register an increase in bike traffic. Who wants to ride one of these clunkers from Freemont to Downtown when you can take the bus? These bikes solve the last mile problem and the tourist needs a bike hop problem. Ie, retail downtown to the Seattle Center, Ride across UW Campus to the U-village etc.
Well, one of the surprising stats is the average trip being 3 miles. That’s a long way – roughly Fremont to downtown. Certainly more than a “last mile” ride.
And the average trip takes 30 minutes. 3 miles in 30 minutes is about 6 mph.
I’ve taken the “clunker” bikes from Fremont to the U-district several times, and it is nearly always faster than than riding the bus if you include wait time – especially on Sundays, when the #32 bus only runs every 30 minutes, while the entire bike ride takes about 10 minutes, even at just 10 mph. Even if my actual destination is up the hill a bit, I can just return the bike at the bottom of the hill next to the Burke-Gilman trail and walk up the hill with my own feet, without the weight of a bike to drag up with me.
Fremont->downtown, I haven’t done on a bikeshare bike, but it’s similar distance to Fremont->U-district and about equally flat. The big difference between the two corridors is probably bike infrastructure. Fremont->U-district, I am willing to ride on the Burke-Gilman trail at 10 mph, during the daytime, without a helmet. Streets of Belltown with door-zone bike lanes and a few mixed-traffic blocks – not so much. Yes, the only cliche is true that a bike route is only as comfortable as the least comfortable section.
I like “app taxis”. I mean, I don’t *like* them, but the term is more appropriate.
I’ve shared a car before (& our family “shares” one now), and profits didn’t go to a corporation.
These bike share companies need about 4 rides per bike per day just to break even. The figures given here come to about 0.3 rides per bike per day, or about 1/13 the number needed to break even.
The figures show that each bike is ridden once (for 30 minutes) about every 3 days. There are several times as many of these bikes in Seattle now as there should be to make any sense whatsoever. Each bike is taking in only about 30 cents per day in revenue. These companies have to be losing money hand over fist so far in Seattle.
The point about free rides is a good one. A guy I know just signed up for Lime Bikes and was given 3 free rides as an incentive to sign up. So even at only 1 ride per bike every 3 days, a lot of those trips are probably free promotional trips.
I think the private companies are lot more patient that Pronto was about being profitable. Pronto was budgeted with the expectation that it would start paying for itself immediately, and when this unrealistic expectation was not met, it immediately ran out of the funds. These companies seem to be playing the long game – build up the user base now, worry about actually making money later.
Where do you get 4 rides per day as a financial sustainability threshold?
https://www.citylab.com/transportation/2017/06/cities-cant-afford-to-let-rogue-bike-shares-run-wild/530664/
A CityLab article.
About the Author
Kate Fillin-Yeh
Kate Fillin-Yeh is the director of strategy at the National Association of City Transportation Officials, working on bike share, safety, and other policy issues.
“NACTO’s estimate1 suggests that a 2,000-bike system with bare-bones staffing would need to rack up more than 4 rides per bike per day in order to break even. This is well above the current U.S. bike-share average. ”
“And the bikes will break: In one conversation with a rogue bike-share company, a city official was told that the company offered two bike models, the “good” 2-year life-expectancy option and another model expected to last less than a year. Can a business model be sustainable with a full system replacement every two years?”
?The model assumed systems that operated 365 days per year at $1/ride. Capital costs were estimated at $200/bike, replaced every 2 years. For operating costs, we assumed 8 staff would be required for under 200 bikes, with one additional staff person for every additional 50 bikes. Staff would be paid $16/hour for 8 hour days with 28% for benefits. Overhead on operating costs was assumed to be 15%.”
Thanks for the source.
Fillin-Yeh’s labor estimate of 1 staff per additional 50 bikes sounds way too high. Limebike and Spin already have 3000+ bikes each. She is probably applying legacy bikeshare standards to dockless, when the model is quite different – bike access through abundance vs access through rebalancing scarcity.
Limebike staff says break-even is at one ride per day:
https://techcrunch.com/2017/12/11/limebike-the-cycle-sharing-startup-expands-to-europe-as-it-hits-1m-rides-in-the-us/
To make sense of it, though, we can compare with Car2Go and ReachNow. These systems have to oversupply to produce and capture latent demand for bike trips. Once they become an institution, they create behavior changes about the way people travel and make land use choices.
The bike utilization rate may be low, but 2,711 trips per day is likely not just a shift from a personal bike to a shared bike. This looks to be capturing new riders, anecdotally confirmed to me by several people I could never convince to bike anywhere who have all become regular bike share riders.
The big test will be winter – if these programs keep more bike riders on the roads than usual, it will be a fantastic impetus for the city to get serious about accommodating it as a major travel mode.
Car2Go and ReachNow might be willing to operate at a loss, under the theory that the exposure will boost sales. Basically, it’s a way to go after the millennials who don’t want a car right now, but might want one someday in the future, when they have kids, and their incomes presumably rise to the point where a luxury car like a BMW actually becomes financially feasible.
Bike share, on the other hand, doesn’t have a “sales” division to subsidize the business, since the bikes that are used for bike share are very different from the bikes that are sold to individuals at bike shops (a necessary consequence of bikes that are constantly exposed outside to weather and thieves).
I’ve worked in other recreation rental businesses where the conventional standard for profitable equipment rental was to charge 1/10 of the initial cost to the business as the per day of rental fee. It’s very hard to understand how this bike share plan can be profitable with these published numbers.
Perhaps a better comparison would be with bike rental shops. According to CBC prices range from 9 / hour to 45 / day. I would say that maxes out at about 5% of the bike cost.
Still, that’s a long way from $1 for 30 minutes. I’m not sure the share companies need 4 rides per day to break even but I am certain they need more than 0.3 rides per day, or whatever the current rate is.
Assuming those bikes cost $500 each (I really have no idea), that means 500 rides just to pay for the bike, or about 4 years. The bikes won’t last 4 years without substantial maintenance and there’s other overhead, of course.
Prove me wrong !
Their ratio of labor and rent to bikes is a lot lower than a bike rental shop.
They’re probably not going to be profitable cumulatively for the first year, because of the initial procurement of the bike.
As time continues, how well the bikes get maintained is going to be key. Anecdotally, I’ve had several instances of problems like seatposts stuck in the lowest position and gear shifters stuck in the highest position. The companies finally added the ability to report a problem with just a few clicks to the app, but they don’t have the system in place to encourage users to actually bother to report problems they find (pure altruism, alone, I don’t think is going to cut it).
One thing I would like to see the companies do is have a deal where if you end your trip within two minutes of starting it, and report a problem, your ride is free. At a minimum, this would give users the opportunity to go try another bike, rather than hobble along with a broken shifter or seat post simply to avoid paying another dollar. In practice, I don’t see people abusing the system by switching perfectly good bikes every 2 minutes, just to save $1 – too much effort for too little money.
Another problem I’m seeing, which probably doesn’t help the maintenance story, is lots of bikes getting knocked over by the wind. Not sure how much better kickstands would help.
Agreed on the ability to report and not have to charge another $. I would have done that many times.
Other fixes for usability (and ideally, financial sustainability):
– Elevation bonus to save re-balancing costs
– Imbalance bonus by incentivizing people to drop off at high demand areas that are showing low supply
– The bike corrals will need to be really clear from the street (don’t rely on people to look in the app to find out where exactly to park. Flags, something.)
– Stopover ability (run into a store and hold a bike so you can use your full 30 minutes if you want). I’ve had trips that I haven’t used a bike because it was just a walking trip with a lot of stops, but I would have used the bike if I could have stopped over.
– Said above, but more bike corrals. No reason not to take one parking spot per block or the corners as mentioned above. That would really help move many of the bikes (All of those that the companies stack up when rebalancing) so that they aren’t in the way and would help alleviate a lot of the angst from people who don’t use them or view them as “clutter” or “garbage” in the sidewalk.
Overall, this entire system will really have the effect of taking car trips off the road when light rail is in place 2024 to distance locations and the bike share serves as a clear last mile connection. Right now, buses have enough stops downtown that it’s not necessarily worth it, but the limited tunnel stops for more people than today may increase that need substantially.
I think the jury is still out on what trips are converting to bike share (previous bike trips, walking trips, Uber/Lyft, etc.)
My hunch is it’s a lot of former walking trips, but at least walking trips that aren’t becoming uber/lyft.
Those are great suggestions, by ja and asdf. I imagine they’ll implement them. They’re probably busy right now with basic functionality, but I’m optimistic that competition will elicit refinement.
The Chinese brands have a credit scoring system where they give points for good behavior and subtract them for bad behavior.
Re: corrals, they could have a ride bonus for using them, e.g. every 10th ride free for each time you park at a corral (as verified by GPS, if they’re accurate enough).
Of has such a credit scoring system. But, without tangible rewards for having a good score (e.g. discounts), it’s mostly meaningless.
hopefully in the next phase the city will still get data from the bike share companies. I would like to see this report after the end of next summer to really see how it’s working
Ofo, by the way, is free until January 2. For me, this is a great thing because one of the limitations I find on the dockless bike share is the use of the bikes for short rides. I could take a Pronto five or six blocks, if there were docks in both directions, because I had the annual pass. It’s silly to do that if it costs a dollar each way. But I’ll be using Ofo these next couple weeks to see if I can get in some short rides between places downtown, say to get to lunch and then back. Perhaps a pricing structure that allowed you to dock for as lunch as it takes to buy lunch on the same ride would make sense.
Excellent news.
You may want to add a clarification that the number of bikes is current, while the trip count is cumulative, so that the “trips per bike per day” cannot be calculated from the numbers supplied, because the bike count varied from 500 to 9400.
What period does this data cover? Your last blog post containing data, in September, mentioned 118,240 rides. If you calculate the difference in days and rides, and assume an average of 6,000 bikes over the Sep-Dec period, you get around 0.4 rides per bike per day, which sounds way too low given that the bikes had 2.56 rides per day per bike in September.
As long as the rides per bike per day is at a financially sustainable level (which I’d guess is 1-2 per day, annualized), it’s not an issue. More bikes tends to mean greater proximity to users, making bikeshare trip times faster; also, peer to peer bike rebalancing (a la Citibike’s Bike Angels) and corrals could increase bike utilization and/or reduce operating costs.
Regarding corrals, they will be an excellent addition. Having say one fixed place per block face in high ridership areas will make locating bikes easier and perhaps promote competition by having different firms’ bikes right next to each other. In addition to ameliorating the clutter issues.
I sure hope these companies succeed because the bikes are being used and this city desperately needs multimodal transportation options. It’s too bad that so much abuse is heaped on these bikes. Why does nobody beat the shit out of all the car2gos haphazardly parked all over the city?
A bike is, by nature, easier to abuse than a car, simply as a result of being light enough for a human to carry and not being enclosed by tons of steel. The companies do what they can, by designing much tougher bikes for bikeshare than what an individual would typically buy at a bike store. But, they can only do so much.
The other problem is that much of the abuse comes, not from people, but from wind and rain. Most of the bikes you see laying down were probably knocked over by the wind, not by humans (as evidenced by the fact that I see many more bikes on their sides immediately after a strong windstorm). If people could simply return their bikes laying against something, rather than relying entirely upon the kickstand, that would go a long way.
There was a long orderly row of spin bikes on the corner of 2nd and Jackson a couple months back. I was leaving work to head to the bus and I watched a disgruntled looking woman walk by and purposely knock down every single bike. She probably knocked down 8 of them. I hollered what are you doing. She grumbled under her breath and found another few and knocked them over. I have no idea what the motivation was. This is an instance where they were reasonably out of the way for pedestrians.
[…] that began operating late this summer surpassed that figure in a little over three months. Seattle Bike Blog: “Spin, ofo and LimeBike are carrying nearly as many daily riders as both the First Hill […]
A Limebike official said recently that their model is financially sustainable on 1 ride per bike per day. That’s pretty insane. Such a low threshold would make bikeshare viable in many, many places.
It also suggests that in places where ridership is say, 2+ rides per day, prices could go down or bike quality go up.
https://techcrunch.com/2017/12/11/limebike-the-cycle-sharing-startup-expands-to-europe-as-it-hits-1m-rides-in-the-us/
For all those hand-wringing about whether these bike companies can be profitable, I suggest checking out these links:
https://www.crunchbase.com/organization/limebike
https://www.crunchbase.com/organization/ofo
https://www.crunchbase.com/organization/skinny-labs-inc
Ofo has over a billion dollars in funding. The least funded of these companies, spin, has 8 million.
These companies can afford to maintain a few thousand bikes for many years. They are probably spending more money on their technology infratructure than the bikes themselves.
In the words, rides per bike may not be that important of a metric… The total volume of rides across all cities needs to be high enough to justify the expense of their engineering staff.
The way these companies will be successful actually has less to do with who succeeds in Seattle than who can capture larger markets… Seattle is just the test-bed.
[…] limited bikeshare data so far works as a Rorschach test; depending how you look you can see a national bikeshare leader or a fragile system with cracks forming. The number of total rides suggests private bikeshare is […]
The average trip distance of 3.0 miles sounds erroneously high, considering that Limebike, which has some 50%+ of market share, reports a 1 mile average trip for Seattle, and elsewhere.
For the 3 miles to be true, the other providers would have to average 5 miles per trip.
https://www.limebike.com/hubfs/EOY%20Data%20Report.pdf
[tweeted this at the author, but also for others looking at the data, to take note]
[…] the bikes after its earlier attempt at a station-based system fizzled. Usage of the dockless bikes quickly eclipsed Seattle’s old system, but the number of trips per bike per day has fallen recently, and remains well below the figure […]
[…] the bikes after its earlier attempt at a station-based system fizzled. Usage of the dockless bikes quickly eclipsed Seattle’s old system, but the number of trips per bike per day has fallen recently, and remains well below the figure […]