A large public bike share is doing well at 4 to 8 trips per bike per day: that is the target in ITDP’s Bikeshare Planning Guide, and NACTO’s figures for September 2023 put the busiest North American systems at 6 to 8. Small networks sit far lower in the published record, and for a residential, campus or neighborhood network that can be fine. Track trips per bike per day, but judge a small network mainly on how many eligible people sign up and ride, how often they ride, and whether a bike is there at the busiest hour.
This article explains how the number is calculated, why the main sources define it differently, which benchmarks exist and where they come from, and when a high or low number is bad news. It ends with the measures and targets that suit a network of one to three stations, and a one-page scorecard for a board or council.
What trips per bike per day measures
Trips per bike per day is the number of trips in a period divided by the number of bikes and by the number of days in that period. It measures how hard each bike works, which is why operators and cities use it to judge both demand and fleet size.
The arithmetic is simple:
Trips per bike per day = trips in the period ÷ (bikes × days)
For example, suppose a hypothetical network logs 420 trips in a 30-day month on 14 bikes. That is 420 ÷ (14 × 30) = 420 ÷ 420 = 1.0 trips per bike per day.
Three choices hide inside the formula, and each can move the result by more than the difference between a good month and a bad one:
- What counts as a trip. Decide whether to count a ride that starts and ends at the same bay within a minute; a rider who undocks and redocks has not gone anywhere. Rebalancing moves by staff and test rides after a repair are not trips.
- Which bikes count. Every bike the program owns, the bikes deployed on the street, or only the bikes available to ride? A bike waiting for a part is owned but not available.
- Which days count. Calendar days, or only the days the system was open? A program that stores its bikes from November to April has a very different annual figure depending on the choice.
A useful variant replaces bikes × days with bike-days available: the sum, over every day, of the bikes that were actually in service. It is the fairest version for a small fleet, where one bike in the shop is a large share of the total.
Related ratios that are easy to confuse
Several other ratios sit close to trips per bike per day, and reports sometimes mix them up:
- Docks per bike compares station capacity with the fleet. ITDP’s guide says “most medium and large systems have 2-2.5 docking stations for each bike in service,” so a returning rider finds a free dock. It describes capacity, not use, and ITDP notes it “is not necessarily used for analyzing system performance.”
- Trips per station per day shows which locations work. It is useful for siting, but a station with more bikes will naturally log more trips.
- Trips per member or per active rider measures habit rather than fleet use. A network can have low trips per bike and high trips per rider if it simply has more bikes than its riders need.
- Bikes per 1,000 residents is ITDP’s sizing guideline, with a planning range of 10 to 30 for large, dense cities. It says how big a fleet is, not how well it is used.
Keep each ratio labeled with its own name in a report, so nobody compares trips per station with trips per bike by accident.
Why sources define it differently
The major sources measure “trips per bike per day” in different ways, so their numbers cannot be compared directly. The difference is almost always in the denominator: which bikes and which days are counted.
| Source | Period covered | Trips counted | Bikes counted | Notes |
|---|---|---|---|---|
| ITDP, Bikeshare Planning Guide (online) | A planning target; city examples are 2017 figures | Average daily trips | Bikes "in active circulation": in a dock, locked or in use, not bikes in repair or the reserve fleet | Target of 4 to 8 daily uses per bike |
| NACTO, report on 2023 (July 2024) | September 2023 | Trips reported by each city | "Available" devices, per city reports | Counts only systems with over 150 vehicles; notes "variance" in how companies and regulators calculate the figure |
| Médard de Chardon, Caruso and Thomas (2017) | 75 systems, published 2017 | Daily trips estimated from publicly available data | Bikes in each system | A third of systems under one trip per bike per day |
| NABSA, 2025 State of the Industry charts | Calendar year 2025 | Annual trips by mode | Vehicles deployed on an average day | NABSA publishes the totals; any per-bike ratio is the reader's arithmetic |
| City permits, per NACTO's 2019 guidelines | Set in each permit | Rides per vehicle per day | Defined by each city | NACTO asks cities to say whether inoperable or damaged vehicles count |
ITDP’s guide is the most explicit about its denominator. It counts “the number of bikes in active circulation,” which “includes bikes either in a dock, locked, or in use,” and excludes bikes “being repaired or … part of the contingency fleet.” NACTO’s report on 2023 uses “trips taken per available device per day” and states plainly in its methodology: “There is variance in how shared micromobility companies and regulators calculate rides per vehicle per day. For the purposes of this analysis, city specific utilization was reported by individual cities.” NACTO’s 2019 permit guidelines make the same point from the regulator’s side, asking cities to decide whether a fleet count includes vehicles that are “inoperable, damaged, unsafe, or in maintenance, or only vehicles that are currently available.”
Small choices move small numbers a lot. In the shared fictional example used across this site, Larkspur Commons, week 5 had 104 rides on 14 bikes, with one bike out of service for 3 days after a flat tire:
Example (fictional). Larkspur Commons, week 5: 104 rides. Counting all 14 bikes for 7 days gives 98 bike-days and 104 ÷ 98 = 1.06 trips per bike per day. Counting only bike-days in service gives 98 minus 3 = 95 bike-days and 104 ÷ 95 = 1.09. Same week, same rides, two defensible answers. If the program then closed for winter and the figure were averaged over a whole calendar year, the number would fall further without any rider changing behavior.
The rule for anyone comparing figures: ask what was counted above and below the line, and over which days, before comparing two numbers.
Five questions before you compare two figures
- Which trips? Are very short rides, staff moves and test rides excluded?
- Which bikes? All owned bikes, bikes deployed, or bikes available to ride?
- Which days? Calendar days, operating days, or a single month such as September?
- Which season? A summer month and an annual average are different measurements.
- Who calculated it? An operator, a city, a researcher estimating from public data, or a reader dividing published totals?
If two figures differ on any of these, treat the comparison as rough, and say so when you present it to a board.
The benchmarks, and where they come from
Published benchmarks fall into three bands: 4 to 8 trips per bike per day for a healthy large public system, 6 to 8 for the busiest systems in North America, and under one for a third of the systems in one global study. Each comes from a different kind of system and a different definition.
ITDP’s target of 4 to 8. ITDP’s planning guide sets “4-8 daily uses per bike” as the target for average daily trips per bike. Below four, it warns, “can result in financial unsustainability for the operator (i.e., user fees not able to cover cost to operate each bike)”, while “more than eight daily uses can indicate limited bike availability, especially during peak hours.” It cites New York City (6.4), Barcelona (6.4), Mexico City (5.4) and Guangzhou (5.0) as showing “solid daily usage numbers in 2017.” ITDP’s 2018 dockless brief adds that Washington’s Capital Bikeshare was “averaging four trips per bike per day” and Dublin’s station-based system 5.6.
NACTO’s busiest systems at 6 to 8. In September 2023, according to NACTO, Bixi in Montreal, Bike Share Toronto and Citi Bike in New York City “were the busiest bike share systems in North America, with each bike in the systems taken out an average of 6 to 8 times per day.” In systems that offer both, NACTO found e-bikes used far more than pedal bikes: nearly 8 times as much in Los Angeles and nearly 4 times in New York City. Yet the Boston area’s Bluebikes, with no e-bikes in 2023, “served more trips per bike per day than almost every other station-based system.” For dockless e-bikes, NACTO’s chart for September 2022 and 2023 shows values between about 0.3 and 2 across eleven cities.
One trip per bike per day as a floor. Médard de Chardon, Caruso and Thomas estimated daily trips from public data for 75 systems around the world and found that a third of them had fewer than “the psychologically important one trip per bicycle per day” (Transportation Research Part A, 2017).
Why big systems score higher. The systems at the top of these lists share conditions a small network cannot copy: dense stations a few minutes’ walk apart, large numbers of residents, workers and visitors within reach, and, in NACTO’s words, “extensive, well-connected networks of protected bike lanes.” NACTO’s report says that “a large number of densely-distributed stations and shared micromobility devices within walking distance of residences, employment hubs, and other popular destinations is critical for supporting trips,” and points to Hamilton, Ontario, where researchers linked the greatest increases in bike kilometers traveled to concrete-protected and well-connected bike lanes, and where bike share trips rose 30% from 2022 to 2023. A three-station network serving one apartment community has none of that scale, so its trips per bike will be lower even when every resident who wants a bike gets one.
NABSA’s 2025 totals. NABSA’s charts for 2025 count 132.2 million trips on docked and hybrid bikes and about 114,000 such bikes deployed on an average day, plus 13.6 million trips on about 46,000 dockless bikes. Divided by 365 days, that is about 3.2 trips per deployed docked or hybrid bike per day and about 0.8 per dockless bike. By the same arithmetic, shared e-bikes (74.2 million trips, 72,000 bikes) come to about 2.8 and shared pedal bikes (71.6 million trips, 88,000 bikes) about 2.2. These ratios are our arithmetic, averaged over the calendar year including seasonal closures, not figures NABSA publishes.
What the small-system record shows
Published figures for small systems are rare, and the few that exist sit far below the big-city bands. They also show how easily a small program’s numbers go unreported, because the denominator is often missing.
- Kailua, Hawaii. A two-station pilot with 12 bikes averaged 41.5 rides a month from 2011 to 2014 (Choy and colleagues, 2015). That is roughly 0.11 trips per bike per day by our arithmetic, and about 0.03 in 2014, when rides fell to 10.6 a month.
- Spartanburg, South Carolina. The same paper cites a system with two stations and 14 bikes that averaged nine rides a day, or about 0.64 trips per bike per day by our arithmetic.
- Portsmouth, New Hampshire. The city reported 941 sign-ups and 1,876 trips from May 1 to November 27, 2017, an average of 63 trips a week. It did not state the number of bikes, so trips per bike per day cannot be calculated; the figure it does allow is about two trips per sign-up over the season.
- Minneapolis, a winter pilot. In a winter extension for a 300-scooter fleet, city staff told the Better Bike Share Partnership they had expected 0.5 to 0.75 trips per vehicle per day and saw about 1.2 to 1.3. Setting a season-specific expectation in advance is exactly what a small network should do.
Two lessons follow. First, small systems need their own yardstick, because a big-city target would label nearly every one of them a failure. Second, the number is noisy at small scale: in a 14-bike fleet, one bike in the shop is 7% of the denominator, and one rainy week can halve the numerator. Read it as a trend over several weeks, not a verdict on one.
Why a high number can be bad news
A high trips-per-bike figure means each bike is busy, and past a point that means riders are arriving to empty docks. ITDP’s guide says it directly: “A high number of average daily trips per bike may indicate that there are too few bicycles in circulation,” and station-based systems in that position should add stations and bikes.
For a small network, the problem appears well below ITDP’s ceiling of eight, because demand piles up in time and place. Residents leave for work in the same hour, and students head to the same classroom building. A rider who finds an empty station at 7:45 a.m. does not show up in the average; that rider simply does not ride, and may not try again.
That is why sizing tools plan for the busiest day rather than the average. The fleet size calculator on this site asks how many rides one bike can serve on the busiest day, with a default of 2. In its fictional Larkspur Commons default, 27 rides on the busiest day ÷ 2 rides per bike = 13.5, rounded up to 14 bikes. The same plan expects about 19.3 rides on an average day, so the network is designed to run at about 19.3 ÷ 14 = 1.38 trips per bike per day on average and about 27 ÷ 14 = 1.93 on its busiest day. A small network that is working as designed sits far below the big-city band.
The fictional Larkspur weeks show the point. In weeks 6 to 8, at only 1.1 to 1.3 trips per bike per day, one station ran empty on weekday mornings. The fix in the example was rebalancing, moving 2 bikes to that station each evening, not buying more bikes. A board looking only at the average would have seen a quiet network; the availability field in the report showed riders being turned away.
Why a low number is not always failure
A low trips-per-bike figure is a problem for a fee-funded public system and a question, not a verdict, for an amenity network. ITDP’s floor of four is explicitly about money: it is the point where “user fees [are] not able to cover cost to operate each bike.” A resident amenity paid for by a property or an HOA is not funded by fees, so the test is different: what the program costs per trip and per resident served, set against what the host wants from it.
Médard de Chardon and colleagues call one trip per bike per day “psychologically important”, a threshold of perception rather than of finance. It matters for public systems that must justify street space and public money. For a station on private land serving 450 residents, the more relevant question is whether those residents can count on a bike.
ITDP itself treats a low number as a diagnosis to make, not a conclusion. Its guide notes that few uses per bike “might indicate inefficient usage of infrastructure and low cost-benefit, likely because of a surplus of bikes,” and suggests two responses: shrink or consolidate stations, or “bolster education and marketing campaigns” when the barrier is awareness rather than infrastructure. The table below turns that into a checklist for a small network.
| What you see | Likely cause | What to check first |
|---|---|---|
| Low trips per bike, few members | People do not know about the program or cannot sign up easily | Sign-up share, the sign-up steps, launch materials |
| Low trips per bike, many members who rarely ride | The station is in the wrong place or the bikes do not suit the trips | Where members live and go; the station’s location against the busiest door |
| Low at one station, high at another | Demand sits on one side of the community | Rebalancing moves; whether to move bikes or a station |
| Low in one month only | Weather, holidays or bikes out of service | Out-of-service bike-days; the same month last year |
| Low across the board in winter | Season | Whether a reduced winter fleet or a winter storage period fits better |
| Falling over several months | Reliability, such as soft brakes, flat tires or empty batteries | Issues and fixes; repeat problems with the same bikes |
Cost per trip as the amenity test
For a host that pays a fixed program cost, cost per trip is the number that connects use to money: program cost for the period ÷ trips in the period. It falls as riding grows, and it makes the trade-off in fleet size visible. Adding bikes raises the program cost and, if riders were not being turned away, barely raises trips, so cost per trip rises. Adding bikes where a station runs empty every morning raises trips as well, and cost per trip may hold steady or fall.
A host does not need a benchmark from another city to use this. Compare cost per trip month by month, set it against what the program is meant to do (a resident amenity, a parking alternative, a campus service), and decide whether that price per trip is worth paying.
How fleet size moves the number
Trips per bike per day depends as much on how many bikes you buy as on how many people ride. With the same riders, a smaller fleet scores higher and runs out more often; a larger fleet scores lower and waits at the station more often. ITDP’s guide makes the same point for large systems: its ratio of bikes to residents “should be large enough to meet demand, but not so large as to have fewer than four daily uses per bike.”
Example (fictional). Hold the Larkspur Commons plan’s riders constant: about 19.3 rides on an average day and 27 on the busiest day. Then change only the number of bikes. The table below shows what happens to the average trips per bike per day and to the rides each bike must serve on the busiest day. The plan’s choice of 14 bikes comes from the fleet size calculator’s design value of 2 rides per bike on the busiest day.
| Bikes | Average-day trips per bike | Busiest-day rides per bike | What a host would likely notice |
|---|---|---|---|
| 10 | 1.93 | 2.70 | Busy bikes and the highest score; empty stations likely at peak |
| 12 | 1.61 | 2.25 | Fewer bikes waiting; some peak shortfalls |
| 14 (plan) | 1.38 | 1.93 | The plan’s balance between availability and cost |
| 16 | 1.21 | 1.69 | More bikes waiting in bays; more charging load |
| 18 | 1.07 | 1.50 | The lowest score; bikes rarely run out |
The figures are arithmetic on the fictional plan: 19.3 or 27 rides divided by the number of bikes. None of the rows is “right” in general. A host that values a bike always being there will accept a lower score; a host watching costs closely will accept a few empty mornings. The point is to choose on purpose and to read the trips-per-bike figure in the light of that choice.
Cities that permit dockless fleets use the number the other way round. NACTO’s 2019 guidelines describe “dynamic fleet caps” that let an operator add vehicles when it meets a rides-per-vehicle-per-day target and shrink its fleet when it does not. The fleet size calculator lets you test the same trade-off for your own community before launch.
Better measures for a residential, campus or neighborhood network
A small network should report trips per bike per day alongside measures of reach, frequency and availability. Together they answer the questions a property manager, a campus office or a board actually asks: are people using it, are they coming back, and is a bike there when they need it?
ITDP already pairs its per-bike target with a second measure, “average daily trips per 1,000 residents (in service area),” which it calls a measure of market penetration and says should improve “over time.” Adapting that idea to a community whose eligible riders are known, by name, in a lease, an HOA roll or a housing list, gives a stronger set:
| Measure | How to calculate | What it tells you |
|---|---|---|
| Sign-up share | Members ÷ eligible people | Reach: how much of the community the program serves |
| Active riders | Members who took at least one trip in the month | Whether sign-ups turn into riding |
| Trips per active rider | Trips ÷ active riders, per month | Habit: whether riders come back |
| Trips per 1,000 eligible people per day | Daily trips ÷ eligible people × 1,000 | ITDP’s market penetration measure, scaled to the community |
| Availability at the busiest hour | Share of peak periods with at least one bike at each station | Reliability where it matters most |
| Free bays at the busiest hour | Share of peak periods with at least one empty bay | Whether returning riders can dock |
| Out-of-service bike-days | Days each bike spent out of service, summed | Maintenance and the real size of the fleet |
| Rebalancing moves | Bikes moved between stations, by direction | Where demand is lopsided |
| Trips per bike per day | Trips ÷ bike-days available | Fleet sizing: too many or too few bikes |
| Cost per trip | Program cost ÷ trips | Value, for a host that pays for the program |
How to measure availability in practice
Availability is the measure most hosts skip and the one riders feel most. A practical method for a docked network:
- Pick the busiest window. For a residential host that is usually weekday mornings; for a campus, the minutes before the first classes. Use the trip data to find it.
- Sample the station status. Most station software records how many bikes and free bays each station has; take a reading at regular intervals through the window.
- Count the readings with at least one bike at each station, and separately the readings with at least one free bay.
- Report the share, by station and by week, next to trips per bike per day.
For example, suppose a hypothetical station is checked 24 times across the mornings of one week and has a bike available in 21 of those readings. Its morning availability that week is 21 ÷ 24 = 87.5%. If the same station’s trips per bike are rising at the same time, riders are starting to miss bikes, and rebalancing or more bays are the next step.
Most of these come straight from a standard monthly report. Colibrino’s report to hosts has four fixed fields (trips, active riders, bike availability, and issues and fixes), with rebalancing moves shown alongside, and a CSV export of the host’s program data at any time. Ride data and reporting shows the format. For cost per trip, the amenity budget worksheet sets a program cost against expected rides.
How to set targets before a pilot
The time to agree what success looks like is before launch, in writing, using the same assumptions that sized the fleet. Then the end-of-season decision rests on numbers both sides accepted in advance.
- Count the eligible people. Adults in an apartment community or HOA, students in a residence hall, staff in an office park.
- Estimate the sign-up share. Survey residents if you can. The fleet size calculator’s default of 12% is a planning assumption, not a measured rate.
- Estimate rides per member per week and how much busier the busiest day is than an average one.
- Size the fleet for the busiest day, at about 2 rides per bike, and add spare bays.
- Work out the expected trips per bike per day: average-day rides ÷ bikes. That is the number the network should approach after a few weeks, not the big-city band.
- Write availability targets in words, such as a bike at each station on most weekday mornings, and say how it will be measured.
- Set the review date and what each result triggers: keep, resize, move or end.
Example (fictional). Cottonwood Hall is a fictional residence hall with 400 students, a 12-minute walk from the main classroom building. Its campus office plans on 15% signing up, so 60 members, riding 3 times a week: 180 rides a week, or about 25.7 on an average day. With the busiest day 1.3 times the average, that is about 33.4 rides; at 2 rides per bike on the busiest day, the plan calls for 17 bikes. Expected trips per bike per day is therefore 25.7 ÷ 17 = about 1.5. The written targets for the end-of-semester review are: at least 12% of residents signed up by week 6; at least half of members riding in each month; a bike at the hall’s station on most weekday mornings; and trips per bike per day between 1 and 2 after the first month. If availability fails while trips per bike runs above 2, bikes are added. If trips per bike stays under 1 with good availability, the office runs a second sign-up drive before cutting the fleet. Every number in this example is invented.
The example network walks through the same steps for the fictional Larkspur Commons, from site walk to the day-90 review.
Reading the number month by month
Read trips per bike per day as a trend, per station, against the same season, and always next to availability. A single month’s figure, taken alone, says very little about a small network.
- Expect a ramp. New programs take weeks to reach their level. The fictional Larkspur example rises from about 0.5 in week 1 to about 1.4 by week 12.
- Compare like with like. Compare October with October, not October with July. Leave out days when the system was closed for storage.
- Split by station. A network average can hide one empty station and one full one.
- Note fleet changes. Adding or retiring bikes changes the denominator; record the date.
- Read it with availability. Rising trips per bike with falling availability means riders are starting to miss bikes.
Adjusting for the season
A seasonal network should report its figure over operating days only, and say so. For example, suppose a hypothetical network with 12 bikes runs from April to October, 214 days, and logs 3,200 trips. Over operating days it averages 3,200 ÷ (12 × 214) = about 1.25 trips per bike per day. Spread over all 365 days, the same season reads as about 0.73, a figure that makes a healthy riding season look like a struggling year-round one. Annual totals such as NABSA’s mix both kinds of systems, which is one more reason to keep local figures on operating days.
Simple decision rules keep reviews honest:
- High trips per bike plus empty stations at peak: rebalance first, then add bikes or bays at the station that runs out.
- Low trips per bike with good availability and low sign-up: work on outreach and the sign-up steps before cutting the fleet.
- Low at one station, high at another: move bikes, then consider moving the quiet station.
- Low across the network in the cold months: compare against the written winter plan; a reduced winter fleet or a storage period may fit better. Seasonal operations describes both options.
The same rules apply whatever the model, though dockless systems measure availability by area rather than by station. The docked versus dockless comparison covers how each model reports data, and the small-city bike share record shows what happened to programs that never agreed on their numbers.
A one-page scorecard for your board
A board, council or campus committee needs one page, the same page every month. A short scorecard keeps the conversation on whether the program serves the people it was meant to serve.
- Reach: sign-up share against the target, and new sign-ups this month.
- Use: trips, active riders and trips per active rider, by month.
- Fleet: trips per bike per day per station, with bike-days in service.
- Reliability: availability at the busiest hour, free bays at the busiest hour, and out-of-service bike-days.
- Balance: rebalancing moves, and which station runs empty or full.
- Value: program cost per trip, if the host pays for the program.
- Next decision: the review date and what each result will trigger.
Example (fictional). Here is the scorecard for weeks 5 to 8 at the fictional Larkspur Commons, built only from the shared example’s figures. Rows the example does not supply, such as actual sign-ups, are left as the place a real report would fill in.
| Line | Weeks 5 to 8 (fictional) |
|---|---|
| Reach | Plan assumed 54 members, 12% of about 450 adult residents; actual sign-ups from the account records |
| Use | 460 trips (104, 112, 118 and 126 by week) |
| Fleet | 460 ÷ (14 bikes × 28 days) = 1.17 trips per bike per day; 1.18 on the 389 bike-days in service |
| Reliability | One station ran empty on weekday mornings in weeks 6 to 8 |
| Balance | 2 bikes moved each evening to the station that ran empty, weeks 6 to 8 |
| Issues | One bike out of service for 3 days after a flat tire in week 5 |
| Next decision | The day-90 season review: keep, resize, move or end |
Read together, the lines say the network is busy for its size, one station needs help in the mornings, and the fix so far is rebalancing rather than more bikes.
A scorecard like this goes to the people who decide: the property manager and owner, the HOA board, the campus office, or the downtown organization’s board and its city contact. Keep the format the same every month so trends are visible at a glance.
For a downtown or neighborhood host, the urban neighborhoods page describes the data a board and a city can expect from a pay-per-ride station on private land. More on planning and data for main streets and neighborhoods is in the neighborhood mobility section. If you want to test the numbers for your own site, start with the fleet size calculator, then plan a station so a site walk can replace each assumption with what your doors and riders show.
