Activity-based working (ABW) is easy to sell to a design committee and hard to defend to a CFO. The promise. Give people a range of settings (focus rooms, collaboration zones, quiet booths, social space) instead of one assigned desk each. Sounds modern, but it only earns budget when it is framed as a financial and operational decision, not an interior-design preference. The question leadership actually asks is blunt: if we stop assigning desks and build a richer mix of spaces, what do we get back, and how will we know?
The honest answer is that ABW ROI is measurable, but only if you instrument it before you demolish anything. Most fit-out projects skip the baseline, finish the build, and then have no way to prove the floor is performing. This piece lays out the metrics that survive a board conversation, how to baseline them, and how to run a post-occupancy evaluation that tells you whether the money worked.
Cost per seat vs cost per employee. The number leadership confuses
The single most useful reframing in an ABW business case is the move from cost per seat to cost per employee. They are not the same number, and conflating them is where most ROI arguments fall apart.
Cost per seat is the all-in annual cost of one physical position. Rent, fit-out amortisation, MEP and HVAC running cost, facilities, cleaning. Divided by the number of seats you built. Cost per employee divides that same total cost base by headcount. In a traditional one-desk-per-person office those two numbers are nearly identical. In an ABW office with a sharing ratio (say 8 desks per 10 people, written 0.8:1), you build fewer seats, so cost per seat may rise slightly. Better chairs, more meeting rooms, acoustic treatment. While cost per employee falls because the expensive line item, leased area, shrinks.
Leadership cares about cost per employee because that is what scales with hiring. The discipline is to model both, state your target sharing ratio explicitly, and show the area you are not leasing as the real saving. On Mumbai rents in BKC or Lower Parel, every square foot you avoid leasing is a recurring annual number, not a one-time fit-out line. Which is why the ratio matters more than the finish budget.
Space utilisation. The baseline that justifies the ratio
You cannot defend a sharing ratio you guessed. Space utilisation is the evidence, and it has two distinct measures leadership should never see merged into one figure:
- Occupancy. Are people in the building at all? Badge swipes or Wi-Fi association counts answer this. A typical post-pandemic Indian office runs 50–65% average occupancy across a week, with sharp Tuesday-to-Thursday peaks.
- Utilisation. When people are in, are the desks and rooms actually in use? Desk sensors, room-booking data versus actual presence, and periodic physical observation studies answer this. The gap between "booked" and "used" for meeting rooms is often the most damning data point on the floor.
Run the baseline for at least four to six weeks before design freezes, covering a full work-cycle including a month-end. Capture peak-day utilisation, not just the average. You size the floor for the busy Wednesday, not the quiet Friday. This baseline is the input to a proper workplace strategy exercise, and it is what converts a sharing ratio from an assumption into a defensible design brief.
Real-estate savings. The headline number, done conservatively
Real-estate saving is the line the board remembers, so model it conservatively or it will not be believed. The mechanism is simple: a sharing ratio below 1:1 means you need less area, which means a smaller lease, a smaller fit-out, and lower running costs for the life of the lease.
Be careful with two traps. First, ABW reinvests some of the saved desk area into collaboration and amenity space. You do not bank the full reduction. Second, the saving is realised only at a lease event (a renewal, a consolidation, or a move); if you are mid-term in BKC with three years left, the saving is paper until you act on it. Frame it as area avoided at the next lease decision, and pair it against the fit-out spend so the payback period is explicit. Our office fit-out cost guide gives the per-square-foot ranges you need to make that payback math honest rather than aspirational.
| Metric | What it measures | How to baseline / measure | What good looks like |
|---|---|---|---|
| Cost per employee | Total occupancy cost ÷ headcount | Finance + lease + amortised fit-out + facilities, annual | Falls vs pre-ABW baseline |
| Sharing ratio | Desks built ÷ people supported | Set from peak-day utilisation study | 0.7–0.85:1 for most knowledge teams |
| Average occupancy | People present vs capacity | Badge / Wi-Fi counts over 4–6 weeks | Peak day ≤ 90% of seats built |
| Desk utilisation | Desks in use when occupied | Sensors + observation study | Sustained 60–75% at peak |
| Meeting-room utilisation | Booked vs actually used | Room panel data vs presence sensors | Booked-not-used < 15% |
| Employee experience | Satisfaction with workplace | Pre/post survey + eNPS workplace items | Improves post-occupancy |
| Retention proxy | Regretted attrition trend | HR data, controlled for market | No worsening; ideally improves |
Employee experience and retention. The metric that protects the saving
The fastest way to destroy ABW ROI is to win the real-estate saving and lose the people. If a sharing ratio is set too aggressively, the floor feels like musical chairs, focus work becomes impossible, and the saving is quietly repaid in attrition and lost productivity. So experience metrics are not a soft addendum. They are the guardrail on the financial case.
Measure experience as deliberately as you measure cost. Run a workplace satisfaction survey before the change and at 8–12 weeks after move-in, using the same questions so the comparison is valid. Track a small set of items that ABW directly affects: ability to find a suitable space for the task, ease of locating colleagues, perceived noise and focus, and overall workplace satisfaction. Pair this with HR's regretted-attrition trend, controlled against the broader hiring market so you are not crediting ABW for a slow quarter. The point is not a single hero number; it is a defensible story that the saving was achieved without degrading the experience that retention depends on.
Collaboration outcomes. Measuring what ABW is supposed to improve
ABW is sold partly on collaboration, which is the hardest thing to measure and the easiest to fake. Avoid vanity metrics like "number of collaboration zones built." Instead, instrument behaviour and proxies that leadership recognises:
- Setting-mix utilisation. Are the focus rooms, booths, and collaboration zones actually used in the proportions you designed, or did you over-build social space and under-build quiet space? Sensor data tells you within a month and lets you re-balance before it calcifies.
- Cross-team adjacency. Track whether teams that should interact are physically using shared zones, via anonymised badge or booking patterns. This is where neighbourhood-based ABW layouts earn their keep for larger floors and GCC enterprise fit-outs where multiple functions share a campus.
- Qualitative signal. Short pulse questions on whether people can collaborate when they need to and concentrate when they need to. The two must move together; a floor that improves collaboration while wrecking focus has not succeeded.
None of these are perfect, and you should not pretend they are. State them as directional indicators, triangulated, rather than precise causal proof. That honesty is exactly what makes the rest of your numbers credible to a sceptical board.
Running the post-occupancy evaluation
The post-occupancy evaluation (POE) is where ROI is proven or quietly forgotten. Treat it as a scheduled deliverable, not an afterthought, and build the measurement plan during design. Sensors, survey instruments, and badge-data access are far cheaper to specify in the turnkey design-build scope than to retrofit later.
A workable POE cadence:
- Settle-in (weeks 0–4): resolve snagging and operational teething. Wayfinding, booking-system glitches, acoustic complaints. Do not measure ROI yet; people are still adjusting and the data is noisy.
- First read (weeks 8–12): first utilisation pull and post-move survey against baseline. Re-balance the setting mix if sensors show clear over- or under-provision.
- Steady state (months 6–12): the read that goes to leadership. Cost per employee, realised or projected real-estate saving, utilisation at peak, experience delta, retention trend. This is the number that justifies repeating the model on the next floor.
Done this way, ABW stops being an act of faith. You walk into the board with a before-and-after on the metrics they already use to run the business, and the design choices defend themselves.
Frequently asked questions
There is no universal number. It must come from your own peak-day utilisation study. As a starting reference, many knowledge-work teams sustain 0.7–0.85 desks per person, but client-facing or shift-based teams differ sharply. Set it from four to six weeks of baseline data covering a month-end, size for the peak day, and leave headroom rather than optimising to the average.
Do not trust any reading inside the first month. People are still adjusting and snagging is ongoing. Take a first directional read at 8–12 weeks against your pre-move baseline, and treat the 6–12 month steady-state read as the one for leadership. The financial saving itself is only realised at a lease event, so distinguish "projected" from "realised" clearly.
You can still run a post-occupancy study, but you lose the before-and-after that makes the ROI argument airtight. The pragmatic fix is to instrument the current floor now, run a few weeks of utilisation and a satisfaction survey as a retrospective baseline, and use it to tune the setting mix and inform the next lease decision rather than to prove the original spend.