February 16, 2026

What occupancy data can reveal in just 20 working days?

And why 20 working days is a surprisingly powerful window

Shortly: 20 working days is long enough to capture four full weekly cycles of office behavior: day-of-week differences, peak days, and recurring patterns.

Workplace utilization has often been measured in short “snapshots” (often two weeks) in manual studies. However, in hybrid workplaces, variability is the norm. Density (2025) argues that two-week snapshots can be misleading because the “new workplace is not static.”

So, 20 working days isn’t a magic number, but it’s a high-leverage minimum. It’s long enough to see repeatable rhythms, and short enough to move from insight to action quickly.

Let’s clarify what “occupancy data” actually is (and isn’t)

Occupancy data is often treated as a single metric: “How full is the office?”, but the most useful insights come from splitting it into demand, supply, and patterns.

CBRE (2024) defines three terms that are worth using consistently:
1. Office attendance (How many people visit over a time period)
2. Show-up rate (Attendance divided by total headcount)
3. Utilization rate (Attendance divided by office capacity)

This is more than semantics. In hybrid work, two offices can have the same average attendance but very different peak-day experiences – and this changes whether you need more desks, more collaboration spaces, or better booking rules.

7 things you can learn in 20 working days

 1. The true weekly rhythm

The same office can be calm on Mondays, packed on Tuesdays, and empty early on Fridays. Looking at distributions by day of the week often reveals more than a simple time series, as it exposes patterns such as peak frequency and weekday clustering.

What 20 days reveals:

• Do you have “core days” that need peak-time support?
• Are certain weekdays consistently underused? 
• Is attendance at assigned desks too unpredictable?

2. Your peak demand

Space decisions are usually broken by peaks, not by the mean. CBRE’s report highlights that pre-COVID, full-time on-site work commonly yielded an average show-up rate of about 70% with relatively low daily variation. Today, you often have lower averages but bigger spikes, which makes peak-aware planning essential.

What 20 days reveals:

• How big is the gap between an average day and the busiest day?
• On which days does the office feel comfortably busy, and on which days does it feel too crowded?

3. Seat-sharing potential

If your attendance is predictable enough, you can safely share desks. CBRE shares an example where flexible attendance allowed about 1,5 people per desk and about 20 % less space than with strict “core days.”

What 20 days reveal:

• A realistic desk-sharing ratio for each floor or zone
• Which teams can’t easily share desks: e.g., compliance, special equipment, highly collaborative work
• Where people are most likely to struggle to find a desk (busy peak days + too few desks)

4. Hot zones vs. cold zones

Most offices contain two parallel realities: a few overused areas and many underused pockets.

What 20 days reveal:

• With basic zone-level occupancy data, 20 days is enough to estimate heatmaps by hour or day, quiet and pressured areas and candidate areas for conversion. 

5. Meeting room usage
Meeting rooms are notorious for “calendar reality” diverging from “human reality”. Industry benchmarks often target meeting space utilization of 40–60% range during core hours. Still, the higher hidden cost is mismatched demand: the wrong sizes, wrong locations, or rooms blocked by no-shows.

What 20 days reveal:

• Which room sizes are actively used 
• Whether your issue is not enough rooms or not enough small rooms
• No-shows and ghost-bookings

6. The “purpose profile” of your office: focus vs. collaboration
A lot of organizations redesign for “collaboration”, but occupancy data can confirm whether the space mix matches real behavior. CBRE cites micro-level utilization data indicating that collaboration spaces were used 64% more throughout the day than desks, on average.

Even if your numbers differ, 20 days can validate:

• Are collaboration spaces carrying the day?
• Are desks used mainly as touchdown points?
• Do you need more quiet focus space than you thought?

7. Operational wins: cleaning, security, catering, and support
This is where 20-day data often pays back immediately: because services can match actual demand.

After 20 days you can:

• Adjust cleaning schedules based on used zones
• Staff reception and security around real arrival and departure curves
• Tune catering or coffee services to peak hour loads
• Plan maintenance windows around low-occupancy times

A practical 20-day occupancy sprint

Step 1: Define the decisions before the sensors

Pick 2–5 decisions you want the data to inform, e. g.:

• Do we have enough desks on peak days?
• Which meeting room sizes are missing?
• What zones should be converted?

Step 2: Choose data sources intentionally

Common sources:

• People-counting (presence sensors)
• Badge/access data (entry/exit patterns)
• Booking data (desk and room reservations)

Step 3: Build the core metrics set

In 20 days, focus on:

• Peak daily attendance and peak hour
• Weekday distributions
• Utilization by zone and space type
• Meeting rooms: bookings vs. actual presence

Step 4: Output that drives action

Deliverables that decision-makers usually use:

• 1–2 page executive summary (peaks, constraints, opportunities)
• Weekday occupancy distribution charts and heatmaps by hour or day
• A prioritized list of interventions

The honest limits of 20 days (so we don’t overclaim)

20 working days is enough to identify recurring weekly patterns, peak constraints, obvious mismatches in space types and room sizes, and operational quick wins.

But it may not fully capture seasonality, major organizational events (e.g. reorgs, policy changes) and long-term lease decisions where you’d ideally want a more extensive dataset.

So, treat 20 days as a discovery sprint. It’s fast, credible and actionable enough – then extend measurement for strategic changes if needed.