Ask a coworking operator how full they are and you will get a membership number. Sixty desks sold. Walk the floor at three on a Wednesday and count heads and you will get a different number — say forty-eight. Both are true. Neither is occupancy, and the gap between them is where every important decision in this business gets made badly.
That gap is not a problem to be closed. A hub where sold equals seated at all times is a hub that is leaving money on the floor. The problem is not knowing the gap — not knowing whether it is twelve or thirty, whether it is the same twelve every day, and whether it has been widening for four months while you were busy.
Two different truths: memberships sold and seats occupied
A membership is a commercial object. It belongs to a member, it runs on a plan with a billing period and a price, it includes some number of seats or passes, and it renews or lapses on a date. It answers the question: what am I owed this month.
A seat is a physical object. It sits in a space, it has a label, and at any moment it is allocated or it is free. It answers a different question: what is happening in my building right now.
Almost every hub keeps these in separate places — memberships in a billing sheet or an accountant's file, seats in the community manager's head and a whiteboard by the door. The consequence is that no single person can answer questions that need both. Which plan produces the members who actually show up? Are the corporate accounts using the seats they pay for, or are we about to lose a renewal we never saw coming? If we sell four more dedicated desks, does the floor break?
None of those are exotic analytics. They are the day-to-day questions of running a hub, and they are unanswerable when the commercial record and the physical record never meet.
Dedicated desks, hot desks and day passes: three inventories, one floor
The reason this is harder than it looks is that a coworking floor is not one kind of inventory. It is at least three, sharing the same square feet and behaving in completely different ways.
- Dedicated desks — sold to a member, allocated to a specific seat, and unavailable to anyone else whether or not the member is in the building. Predictable revenue, zero elasticity.
- Hot desks — sold as access to a pool rather than a chair. Elastic by design, and the entire point is that you sell more of them than you have chairs.
- Day passes and drop-ins — sold to people who are not members at all. Pure upside, entirely unpredictable, and the category most likely to arrive on the busiest day.
- Meeting rooms and cabins — booked by the hour or the block, often included in a plan and therefore often given away without anybody noticing.
Each of these consumes the same finite floor, which means capacity is not a number you can compute from memberships alone. Sixty desks sold across those four categories can mean a comfortable floor or a queue at the coffee machine, depending entirely on the mix. If your system knows only that sixty exist, it cannot tell you which.
So the seat has to be a real record — registered to a space with a type and capacity, labelled, allocated or released as an explicit action — and every membership has to be attached to a member, so the same object appears in both stories. That is the whole trick. One member record, one seat record, and a relationship between them that a person can look at.
The live seat map as an operational instrument, not a poster
Most seat maps are decoration — a nice floor plan on a wall or a slide in an investor deck, drawn once and never updated. A useful one is the opposite: unattractive, current, and consulted many times a day.
The test is whether your receptionist looks at it before answering a question. When a walk-in asks whether there is a desk for the afternoon, when a member asks to move nearer the window, when a corporate account wants to add two people from Monday — if the answer to any of those comes from memory or from a phone call to the community manager, you do not have a seat map. You have a picture of one.
The states it has to carry are small and boring. Occupied, by whom, from when. Free and sellable now. Held for someone arriving. Out of service because the monitor arm broke. That last state is the one everybody skips and it distorts everything, because a seat you cannot sell should not be counted as availability you failed to fill.
A seat map that nobody checks before answering a customer is not a seat map. It is a drawing.
Footfall: who is actually in the building today
Allocation tells you who has a right to a seat. Check-in tells you who is exercising it. These are wildly different numbers and the second one is the one your operations actually run on.
A check-in and check-out log gives you a live count of who is on the floor now and a history of usage across the week. Almost every practical decision at a hub is downstream of that log. How much coffee and pantry stock to hold on a Tuesday versus a Friday. Whether the community manager is needed at the desk at nine or at eleven. Whether the second meeting room genuinely runs hot or only feels that way because it is the one nearest the entrance.
It also answers a safety question you will eventually be asked, either by a landlord, a fire inspection or an incident. Who was in the building at four o'clock. An operator who cannot answer that is running a floor full of other people's employees on trust and a visitor register that nobody signs on the way out.
And it is the earliest possible warning of a lapsing account. A corporate member with six dedicated desks whose people have checked in twice this month is not a renewal. It is a cancellation with a date on it that you have not been told about yet — and the only reason you would know in time to save it is the footfall log.
Meeting rooms as inventory you are giving away free
Meeting rooms are the most abused asset in a coworking space, and the abuse is usually the operator's own doing. They get bundled into plans as included hours, booked on a shared calendar or a paper diary, and then never reconciled against what the plan actually included.
Three specific leaks follow. Members exceed their included hours and are never billed for the excess, because nobody is counting. Rooms are block-booked at the start of the week and then not used, so the room shows busy and sits empty. And no-shows are never recorded, which means the operator cannot tell the difference between a room that is genuinely in demand and a room that is being hoarded.
The fix is to treat a booking like a small transaction with a lifecycle: booked for a member for a time slot, then rescheduled, checked in, completed, cancelled or marked no-show. The moment no-show is a status rather than a shrug, two things become visible — which members hoard, and how much of your meeting-room capacity is fictional. Usually enough to postpone building another one.
Bill the overage as its own line, separate from membership. Not because the amount is large, but because a member who is never billed for excess hours will never believe the limit is real, and the operator who quietly absorbs it eventually raises everybody's price instead.
Overselling on purpose, and the point where it stops working
Selling more hot desks than you have chairs is not a trick; it is the business model. Nobody comes in every day. Some members are on client sites four days a week. A hub that only ever sells one flexible membership per chair is running an expensive serviced office and charging coworking prices.
The discipline is knowing your own ratio rather than borrowing someone else's. The only way to know it is to hold two series side by side for a couple of months: flexible memberships active, and peak simultaneous occupancy on each day. The gap between them is your real headroom, and it is specific to your building, your city and your member mix. An operator whose members are consultants has a completely different ratio from one whose members are early-stage product teams who sit together five days a week.
The failure mode is not the average. It is the peak. Nobody cancels because the space was full on average; they cancel because on the one Tuesday they brought a client, there was nowhere to sit. So track the peak day and the peak hour, not the mean, and treat the first week where peak touches capacity as the signal — because the complaints will arrive two months after the number does.
Reading utilisation before you decide to expand
Expansion is where this stops being a reporting question and becomes a money question. Another floor is a lease, a deposit, a fit-out and a payroll line, committed against a belief about demand. Most operators form that belief from a full-looking floor and a waiting list of unclear seriousness.
Before signing anything, separate the things that all feel like demand and are not the same thing at all.
- Real capacity pressure — peak simultaneous occupancy repeatedly at or near sellable seats, across several weeks, not one busy fortnight
- Mix pressure — you are full of the wrong thing; plenty of hot-desk room and no cabins, which is a layout problem that a floor of the same layout will not solve
- Pipeline that is actually converting — enquiries that took a tour and signed, not a list of people who asked a price in January
- Seats that are unsellable rather than unsold — out of service, blocked by a stalled fit-out, or held for a member who never arrived
- Revenue per sellable seat, trending — if it is flat while occupancy climbs, you have been discounting your way to a full floor and a second one will do the same
The last one deserves the most attention and gets the least. Utilisation and revenue can move in opposite directions for a long time without anybody noticing, because the floor looks busy and the community feels alive. A hub that is 90 per cent occupied at a falling average rate is not ready to expand. It is ready to raise prices, and it needs the seat-level data to know which seats can carry it.
The weekly review that takes ten minutes
Same day each week, same five numbers, written down so you build a series rather than a series of impressions.
- Seats sellable, seats allocated, and peak simultaneous check-ins for the week — three numbers, side by side, in that order
- Memberships renewing in the next thirty days, and for each one, its check-in count this month
- Meeting-room hours booked, hours actually used, and no-shows by member
- Day passes sold, and how many of those people had visited before
- Revenue per sellable seat this month against last, and dues outstanding by member
Do that for six weeks and you will find at least one corporate account quietly on its way out, at least one meeting room whose demand is imaginary, and at least one plan whose members never come and who will therefore renew forever or leave suddenly — which is worth knowing which.
None of this requires a bigger team. It requires that the membership and the seat stop living in two different places, so that one screen can hold both. The operators who scale past one floor are rarely the ones with the nicest interiors. They are the ones who can tell you, without walking the floor, exactly how full they are and exactly how much of that fullness is paying.