Meeting room statistics 2026: sourced numbers on occupanc...

Most "meeting room statistics" pages are a list of round numbers with no dates and no methodology, each one linking to another blog that links to a third that has since deleted the post. The number everyone quotes about ghost bookings has been circulating for years with no visible original study behind it.

So this page has a rule: every figure below comes from a named source with a stated sample, and the source is linked. Where we have done our own arithmetic, it says so. Where a widely-repeated number could not be traced to a real study, it is not here, and there is a section near the end explaining which categories of number in this space deserve suspicion. That section is the most useful part of the page if you are about to put one of these figures in a board deck.

Office and room utilization

The best-evidenced numbers in this category come from JLL's Global Occupancy Planning Benchmark Report 2026, which draws on 84 organisations representing 716 million square feet across North America, Latin America, EMEA and Asia Pacific.

Global office utilization reached 56 percent in 2026, up from 54 percent in 2025 and 49 percent in 2024. Pre-pandemic was 61 percent.
The gap between actual and target utilization narrowed to 18 percentage points, from 25 points in both 2024 and 2025. That narrowing came from a 2-point rise in actual utilization and a 5-point cut in target utilization, which is to say organisations lowered their expectations more than employees raised their attendance.
EMEA is the only region where utilization fell, from 58 percent to 55 percent since 2023. Latin America gained the most, up 10 points over two years.
Technical spaces (labs, manufacturing, data centres) sit at 45 percent utilization against a 72 percent target, and make up nearly 23 percent of managed portfolio space.
Improving space data accuracy is now the number two corporate real estate priority, behind portfolio optimization (cited by 71 percent of respondents). 73 percent of organisations have a data governance programme in place, which JLL treats as the head start on making any of this measurable.

Read the 56 percent figure carefully. It is building-level utilization across large corporate portfolios, not meeting room utilization, and not a number a 40-person office should benchmark itself against directly.

Attendance patterns, which drive room demand

Room demand is not spread evenly. Two independent datasets agree on the shape.

62 percent of organisations now require a fixed number of in-office days, up from 49 percent in both 2024 and 2025, and up from just 28 percent in 2022 (JLL, 2026).
The share of employees attending three to four days per week jumped from 36 percent to 55 percent in a single year, the largest year-over-year shift in three years of JLL benchmark data. Globally, 70 percent are now in three to five days per week.
Kastle Systems' Peak Day Hybrid Index put Tuesday peak occupancy at 64.7 percent across its ten-city sample in early July 2026, down from 66.8 percent the prior week, which was a national record. Eight of the ten cities peaked on Tuesday.
Tuesdays carry 23 percent of the weekly meeting load. Fridays carry 16 percent (Microsoft, based on aggregated Microsoft 365 telemetry).

The practical consequence: a room count that works on average fails on Tuesday. If you size rooms to weekly average demand you will be short two days a week and over-supplied on Friday.

The booked-versus-occupied gap

This is the number everyone wants and the number with the weakest public evidence. Being straight about that is more useful than repeating a figure.

What is well evidenced is the symptom. Gensler's 2026 Global Workplace Survey, based on more than 16,400 office workers across 16 countries, found:

43 percent of employees have cancelled meetings outright because rooms were not available.
65 percent use meeting rooms for individual focus work because quiet space is not available elsewhere.
64 percent take calls in hallways for lack of an appropriate space.
Two-thirds report "hacking" their workspace to compensate for gaps in what the office provides.

Note what the 65 percent figure implies for any sensor-based occupancy measurement: a large share of "occupied" room-hours are one person doing solo work, which is a different problem from a well-attended meeting. A sensor cannot tell you the difference. A booking system cannot either.

If you want your own booked-versus-occupied number, the honest method is a two-week manual audit: walk the floor at fixed times, record booking state from the calendar and headcount from the doorway, and compute the ratio yourself. It is unglamorous and it produces a number you can defend.

Meeting size and shape

Meetings cluster at two people, and at three to six. Steelcase instrumented its own Munich Learning and Innovation Center with sensors from January to March, measured 8am to 4pm each working day, and found meetings occurred most often in groups of three to six or in pairs. Their recommendation from the data was to increase two-person spaces and decrease larger ones.
57 percent of meetings are ad hoc calls without a calendar invite, and 1 in 10 scheduled meetings is booked at the last minute (Microsoft telemetry). Microsoft's own methodology note puts the unscheduled or ad hoc share at 60 percent, based on the top 20 percent of users by meeting volume, which is worth knowing before quoting either number.
Large meetings of 65 or more attendees are the fastest-growing meeting type (Microsoft). Almost all of these are fully remote, so this is not an argument for bigger rooms.
Nearly a third of meetings now span multiple time zones, up 8 percentage points since 2021 (Microsoft).
There are an estimated 32.4 million huddle rooms worldwide, and fewer than 2 percent were video-enabled when Frost & Sullivan published that figure. That was 2018. The number is still quoted as current across the industry, which it is not.

When meetings actually happen

All from Microsoft's aggregated Microsoft 365 telemetry, ending 15 February 2025, excluding education and EU tenants.

Half of all meetings (50 percent) fall between 9am to 11am and 1pm to 3pm. That is a room demand curve, not a flat load, and it is why five rooms can feel like two.
Meetings starting after 8pm are up 16 percent year over year.
PowerPoint edits spike 122 percent in the final 10 minutes before a meeting. This is the mechanism behind late arrivals better than any survey question about punctuality.
Employees are interrupted every 2 minutes during core hours, 275 times a day by meetings, emails or chats (based on the top 20 percent of users by ping volume).
The average worker receives 153 Teams messages per weekday and 117 emails.

What a meeting room costs

Room cost is one place where the underlying data is solid and the derived numbers are easy to compute yourself.

US average asking office rent was USD 37.58 per square foot in Q2 2026, up 2.6 percent year over year, the fastest pace in six years and above the trailing 30-year average (CBRE).
Taking rents ran about 10.1 percent below asking rents in Q2 2026, wider than the 8.6 percent spread in 2019. In Q1 2026 CBRE put asking at USD 37.21 and taking at USD 33.35.
Video-enabling a huddle room costs USD 2,000 to 5,000 depending on device and peripherals (Frost & Sullivan, 2018, so treat as a floor rather than a current price).

The following is our own arithmetic from the CBRE figure, not a cited statistic:

A 150 square foot huddle room at USD 37.58 per square foot is roughly USD 5,600 per year in rent.
A 400 square foot conference room is roughly USD 15,000 per year.
At 250 working days and 8 bookable hours per day, that conference room costs about USD 7.50 per available hour. If it runs at 40 percent real occupancy, the effective cost is closer to USD 19 per used hour.

Run that last calculation with your own rent and your own occupancy before quoting anyone else's version of it.

Which numbers to distrust

This is the section other roundups leave out, and it is the one that will save you an embarrassing slide.

No-show and ghost booking percentages

You will see figures between 30 and 45 percent quoted constantly. Almost every one of them originates from a workplace-analytics or sensor vendor, describing their own customer deployments, without a published sample size, date range or definition of "no-show". Those companies sell the product that fixes the problem the number describes. That does not make the numbers wrong, but it does mean they are self-selected (offices that install occupancy sensors are usually offices that already suspect they have a problem) and unauditable. We have not put a headline no-show percentage on this page for that reason.

Numbers with no date attached

The Frost & Sullivan "32.4 million huddle rooms" figure is from 2018 and still appears in 2026 marketing pages as current. Check the publication date on anything you plan to cite, then check whether the source it cites has one.

Vendor-sponsored survey research generally

JLL, CBRE, Gensler and Steelcase all produce genuinely rigorous research, and all of them sell services that benefit from the conclusions. JLL sells occupancy planning and its report concludes occupancy data matters. Steelcase sells furniture and its study concludes you need different furniture. Gensler designs offices and its survey concludes offices need redesigning. The methodology in these reports is usually disclosed, which is what makes them usable, but read the incentive alongside the number. The same caution applies to us: we sell room displays, and we have an interest in you concluding that room state should be visible.

Anything about "the average meeting length"

Widely quoted, rarely sourced to anything measurable. Calendar duration and actual duration are different quantities, most systems only record the first, and default calendar increments (30 and 60 minutes) shape the data far more than meeting content does. Treat any average meeting length figure as a description of your calendar's default settings.

The general rule

If a statistic does not come with a sample size, a date and a definition, it is a talking point, not a measurement. Your own two-week floor audit beats every number on this page for the purpose of deciding what to do about your own office.

The buy side of this decision

The numbers above mostly point in one direction: room state is hard to see, demand is concentrated on two or three days, and most of the loss is rooms that are booked but not used. Measuring your own version of that requires the booking data (which you already have in Google Workspace or Microsoft 365) and some way of putting it in front of people at the door.

You can build that side yourself. Microsoft Graph or the Google Calendar API, a token refresh loop, a browser in kiosk mode. It is a good weekend project and a mediocre ongoing commitment once you have more than a handful of rooms.

Lobby is the buy side. It syncs in real time with Google Workspace and Microsoft 365 room resources, and shows the same data three ways from one dashboard: a physical e-ink display on TRMNL 7.5" open-source hardware, a virtual display that is just a URL in any browser tab, or a room overview board for a lobby or reception.

The hardware is bought direct from the TRMNL shop and we take no markup on it. Volume pricing applies: USD 139 per device for 1 to 9, USD 119 each from 10 to 20, USD 109 from 21 to 50, and down to USD 99 at 151 or more, at the same list price in the EU and the USA, before tax, duties and shipping. Buy once and own it. Magnetic mount, up to 12 months of battery, readable from 5m.

On the software side, Free is free forever for up to three active displays with 15-minute e-ink updates and no credit card. It is not a trial. Unlimited is USD 30 per month billed yearly, which is USD 360 a year, or USD 50 per month if you pay month to month, for unlimited rooms and displays. Pro Unlimited is USD 60 per month billed yearly, USD 720 a year, or USD 100 per month month to month, and adds 5-minute e-ink updates, custom templates, your own logo with Lobby branding removed, room feedback with free QR stickers, a Slack /book command, priority support, full EU hosting for TRMNL devices, battery saver mode and smart updates. Yearly billing saves 40 percent and there is no annual contract lock-in. Setup takes under 10 minutes, self-serve, no sales call.

And plainly, what it is not: no desk booking, no visitor management, no room automation or AV control, no enterprise SSO or SCIM at scale, no MDM on the TRMNL hardware, no on-premises Exchange without hybrid, and no occupancy sensors. If you want the sensor-derived numbers this post declined to quote, you need a different category of product. Lobby is made by Vikba ApS in Denmark and is not affiliated with Microsoft or Google.

TL;DR

Global office utilization is 56 percent (JLL 2026, 84 organisations, 716 million square feet), up from 49 percent in 2024 and still below the 61 percent pre-pandemic level.
Demand is concentrated: Tuesday carries 23 percent of the meeting load against Friday's 16 percent, and half of all meetings fall in the 9 to 11 and 1 to 3 windows.
43 percent of employees have cancelled a meeting because no room was available, and 65 percent use meeting rooms for solo focus work (Gensler 2026, more than 16,400 workers, 16 countries).
Meetings cluster at two people and at three to six (Steelcase sensor study). 57 percent are ad hoc with no calendar invite (Microsoft telemetry).
US average asking rent is USD 37.58 per square foot (CBRE Q2 2026). By our own arithmetic that puts a 400 square foot conference room at about USD 15,000 per year, or roughly USD 19 per hour actually used at 40 percent occupancy.
Treat every no-show percentage you see with suspicion. Nearly all of them come from vendors measuring their own customers without published methodology. Run a two-week manual audit instead.

Sources

Global Occupancy Planning Benchmark Report 2026, JLL Research, published 19 May 2026. 84 organisations, 716 million square feet.
Breaking down the infinite workday, Microsoft Work Trend Index special report, June 2025. Aggregated Microsoft 365 telemetry ending 15 February 2025, plus a 31,000-respondent survey across 31 markets. Methodology notes are published on the page.
Gensler 2026 Global Workplace Survey, more than 16,400 office workers across 16 countries, and Facility Executive's summary for the 43, 65 and 64 percent breakdowns.
Kastle Systems Peak Day Hybrid Index, badge-swipe data across a ten-city sample.
Using Data to Create a Human-Centered Workplace, Steelcase Workplace Advisor sensor study, Munich LINC, January to March, 8am to 4pm daily.
Q2 2026 U.S. Office Market Report and Q1 2026 U.S. Office Market Report, CBRE, for asking and taking rents.
Huddle-room Collaboration: Looking Beyond the Hype, Frost & Sullivan, February 2018, for the huddle room count and video-enablement cost. Note the date.
Lobby plan and hardware prices were verified on usethelobby.com/pricing and usethelobby.com/hardware on 5 August 2026. Check those pages for the current figures before you budget.

Meeting Room Statistics 2026: Occupancy, No-Shows, Room Sizes and Cost