Someone circled the lot for ten minutes on Tuesday, and now a request for more parking is on your desk. The complaint is real. Whether the lot is too small is a different question, and office parking utilization answers it: five numbers you can count yourself in two weeks, with no sensors.
- Most parking complaints are peak-day complaints: CBRE finds 73% of organizations at 61-100% of capacity on the peak day and 66% below 60% on an average day.
- Attendance data is not parking data. Badge systems count people, work plans count intentions, and neither counts cars.
- Three of the five numbers point at allocation you can change in a week. Only two ever justify capital.
- A structured space is a five-figure decision: WGI puts the 2024 US national median at $29,900 per space.
What is office parking utilization?
Every argument about parking supply is really an argument about which hour you counted.
Office parking utilization is the share of your parking spaces occupied by vehicles at a given moment, measured as cars present divided by spaces provided. It is counted by hour and by day, because a lot that is full at 10am on Wednesday can be half empty at the same hour on Friday.
Some call this the parking occupancy rate; the math is the same either way.
Two things are deliberately missing. A permit is an entitlement, not a car. A badge-in is a person, and people arrive in carpools, on transit and on foot.
Full on Tuesday, empty on Friday, and both are true
Workplace demand is not flat across the week, and that is measured. In its 2025 Americas Office Occupier Sentiment Survey, CBRE found 73% of organizations hitting 61-100% of capacity on their peak day, only 34% at capacity on an average day, and 66% running below 60%.
Kastle Systems, which counts badge-ins across commercial buildings, shows the same shape: in February 2024 the peak day reached 63.1% against a 53.0% weekly average, with the peak midweek, the low on Monday and Friday, and a gap Kastle calls widening.
Neither counts a vehicle: Kastle counts badge swipes and CBRE counts space utilization. They give you the shape of the week, not the cars in your lot.
The five numbers
Compute all five over one two-week window. Three describe how you hand out the spaces you own. Two describe the lot itself, and only those two justify new concrete.
1. Permits or passes issued per space occupied. Divide credentials in circulation by the average vehicles present on a peak day. Issuing more credentials than spaces is a deliberate choice, the way airlines oversell seats, and we found no published benchmark for this ratio from any transportation body. Track your own number quarterly.
2. Peak-day occupancy, by day of week. Your busiest hour on your busiest day, as a share of general spaces. One of the two numbers that can justify building. Planners treat a lot as operationally full near 85%, past which drivers cruise for the last spaces, but that is a rule of thumb: the Institute of Transportation Engineers uses "reasonably full occupancy (i.e., at least 85 percent)" only to decide whether a site suits a parking survey.
3. No-show rate on reserved spots. Spaces booked but empty at the peak hour, over spaces booked. Every point is a space you own and cannot use. No transportation body, parking association or municipal source we checked publishes a workplace no-show benchmark, and a meeting-room figure would not transfer.
4. Recoverable capacity on the peak day. Empty general spaces at the peak hour of the peak day. This is where parking business cases quietly cheat: Friday's empty half of the lot is worth nothing to the person who cannot park on Tuesday.
5. Carve-outs as a share of the lot. Accessible, EV charging, visitor and contractually committed spaces, counted separately and never added to the recoverable pool. The second number that can justify building, because carve-outs only move one way.
How to count cars for two weeks
Nothing in your stack counts vehicles, so for two weeks a person with a phone is the instrument. The protocol below is our recommendation, not an industry standard. ITE sets the parts that decide data quality: tie every observation to the hour of day, use spot counts at fixed intervals, exclude sites with construction underway, and use sites at least two years old.
- Count cars, not credentials
Walk the lot and count occupied spaces. Permits, badge logs and bookings measure something else.
- Fix the clock before the day
Two counts daily, mid-morning and mid-afternoon, at the same clock time. ITE treats the hour of day as critical.
- Cover Tuesday to Thursday for two straight weeks
Six counting days catch the midweek peak twice, and week two checks week one.
- Log carve-outs on their own line
Accessible, EV, visitor and reserved-by-contract spaces get counted, then excluded from the general-pool math.
- Pull entitlements for the same days
Permits in circulation and reservations made. The gap against your counts is metrics one and three.
Skip any week with a holiday, an all-hands or on-site construction, or the business case collapses under the first question.
What Microsoft 365 gives you, and what it cannot
Microsoft Places already runs this comparison in another domain. Places building analytics sets work plans, the days employees say they are coming in, against occupancy inferred from badge data you upload. A good proxy for who will be in the building, not a count of cars, and Microsoft does not claim otherwise.
Copy Microsoft's privacy floors: badge data kept 28 days, aggregates 90 days, nothing shown for leaders with fewer than 10 reports or days with fewer than 10 people in the building. Apply the same rules to parking data.
In Europe this is law. France's CNIL says access-control logs should be deleted three months after recording, and staff representatives are informed or consulted before you install access or working-time monitoring. In the US the laws that bite target cameras, not counting: California's Civil Code section 1798.90.51 requires plate-reader operators to publish a usage and privacy policy covering purposes, retention and destruction, and Illinois BIPA covers six biometric identifiers including face geometry, as summarized by Morrison Foerster. A plate is not one of the six.
When the data really does say build
Allocation fixes are cheap. Building is not.
was the US national median cost to construct one space in a parking structure in 2024, up 3.1% over 2023, in an annual survey built from a database of completed structures.
Source: WGI Parking Structure Cost Outlook for 2024In NAIOP's Development Magazine, John W. Dorsett of Walker Consultants puts a surface space at $3,000 to $8,000 or more and a multilevel structured space at $25,000 to $100,000 or more, plus $150 to $1,000 or more per space each year to operate. Being wrong on asphalt is survivable. Being wrong on a deck is not.
Four cases favor capacity. Overlapping shifts, where peak demand approaches the sum of two headcounts. Sites with no viable transit alternative. Accessible parking, since the 2010 ADA Standards scale the minimum with the spaces you provide, count per parking facility, and require one van-accessible space per six accessible ones. And EV charging: the US Energy Information Administration reports about 22% of US light-duty vehicles sold in 2025 were hybrid, battery electric or plug-in hybrid, up from 20% in 2024, and a charging bay is occupied for the length of a charge.
The trend runs against surplus. JLL's 2026 occupancy benchmark, covering 84 organizations and 716 million square feet, puts global utilization at 56%, up from 54% in 2025 and 49% in 2024. Allocation fixes buy headroom, not a permanently smaller lot.
You will also meet the claim that some precise percentage of employees stay home over parking. We went looking for that survey and could not find it. What is measured sits at employer level: in the same CBRE survey, 52% of occupiers would reject a building lacking parking, 65% among those with a mostly suburban footprint. That is employers ruling out buildings, not employees deciding whether to drive in, and we found no credible measurement of the second thing.
How Parking Pro helps
Counting by hand works for two weeks. Keeping the numbers current forever is a tooling question.
Parking Pro runs workplace parking inside Microsoft Teams, where hybrid staff already plan their week. People book a spot for the days they come in. You keep a fixed core of spaces for the needs that require them and open the rest as a bookable pool.
Two of the five numbers then come as a by-product. Reservations give you the entitlement side of metric one and the denominator of metric three, without a spreadsheet someone rebuilds each quarter. Keep counting cars for the occupancy side, because bookings measure intent and only a count measures presence. If the lot is your recurring problem, get Parking Pro for Microsoft Teams and start with your peak day.
The two-week version
Count cars twice a day on six midweek days, hold carve-outs aside, and compare permits and reservations against what you counted. If peak occupancy stays high after you clear no-shows and reclaim unused reservations, you have a supply problem worth costing. If it drops, you have an allocation problem, and you just saved a five-figure line item per space.
The same method works for rooms: see meeting room utilization. If allocation is where your lot hurts, the fix is a fixed core of reserved spaces plus a bookable pool, not more concrete.
Frequently asked questions
What is a good office parking utilization rate?
We found no published benchmark for corporate lots. Planners commonly treat a facility as operationally full near 85%, and ITE uses at least 85 percent occupancy to test survey suitability. Both are conventions, not pass marks.
How do I measure parking utilization without sensors?
Count occupied spaces by hand, twice a day at fixed clock times, Tuesday to Thursday of two consecutive weeks, logging accessible, EV and visitor spaces separately. ITE ties observations to the hour of day and excludes sites under construction.
What is a normal parking no-show rate?
Publicly unknown. We found no no-show benchmark for workplace parking from any transportation body, parking association or municipal source. Compute your own: reserved spaces empty at the peak hour over reserved spaces, then watch it after you change the rules.
Does a lack of parking reduce office attendance?
Widely asserted, poorly measured. No credible survey quantifies how much parking availability changes an individual's decision to come in. The nearest sourced finding is at employer level: CBRE reports 52% of occupiers would reject a building lacking parking, 65% among suburban-footprint occupiers.
