Robots in Service Charges: Apportioning a Robot Across Tenants
The Building Decides Whether the Landlord Can Buy At All
In a single-occupier building, a robot purchase is a capital decision made by the occupier. In a multi-let building it is a recovery decision made by a landlord who will not keep the saving, and that changes almost everything about how the case has to be built. The landlord pays for the machine, the tenants pay for it through the service charge, and the benefit lands on whichever occupiers happen to be in the areas the machine serves. Unless the lease already allows a charge of that kind, the landlord either absorbs the cost or does not buy.
This is the reason so many multi-tenant deployments stall after a successful pilot: the trial is authorised by the property manager, the machine works, and the purchase fails in the service charge audit rather than in the capital committee. The question that decides it is not whether a robot can clean a lobby. It is whether a robot can be lawfully apportioned across a schedule of tenants, and how.
What follows sets out the three apportionment methods used in practice, worked against a real schedule of floor areas and footfall, together with the lease language each method requires and the challenges each one attracts from a well-advised tenant. It is written for the property manager and the asset manager rather than for the robot supplier, because the supplier's part of the problem is the easy half.
Where a Robot Sits in a Service Charge Budget
Before apportionment there is classification, and classification is where the first argument starts. A service charge budget has conventional cost headings, and a robot can plausibly land in four of them with very different recovery consequences.
| Treatment | Where it appears | Recovery profile | Best used when |
|---|---|---|---|
| Operating expense, outright purchase | Cleaning or security line, depreciated over the asset life | Recovered annually in proportion to the apportionment; capital risk sits with the landlord | The machine is expected to outlast the lease term and the landlord wants the residual value |
| Operating expense, subscription | Cleaning or security line, as a recurring fee | Recovered annually at the contracted fee; no depreciation judgement needed | The building has short lease terms and the cost should track occupancy |
| Reserve or sinking fund | Capital reserve, funded over several years | Recovered through the reserve contribution, smoothing the cost across years | The machine is expensive relative to the annual budget and a level charge is wanted |
| Landlord's own account | Not in the service charge at all | Not recovered | The lease does not permit the charge, or the benefit accrues mainly to the landlord's own marketing of void space |
The distinction between the first two is not academic. An outright purchase inside a service charge creates a depreciation judgement: the machine is written down over five years while the tenants who paid for it may all have left by year three. A subscription converts that into a running cost with no residual-value argument, which is why a landlord with a short average lease term should generally prefer it even at a slightly higher total price. The economics of that trade are the same ones analysed in lease versus buy for service robots.
The reserve treatment matters in buildings with a formal sinking fund: a machine placed in the reserve is charged to tenants over a period longer than its deployment, which is attractive politically and defensible only if the reserve is genuinely for capital replacements rather than a smoothing device. Auditors look at this, and a machine that appears in the reserve one year and in operating expense the next invites a challenge on consistency.
Method One: Apportionment by Net Internal Area
Area apportionment is the default in most leases because it is the easiest to verify. Every lease already states a tenant's floor area, the service charge is expressed as a rate per square metre, and a new cost line simply enters the pool.
Take a 9,400 square metre net internal area building with four tenants and common parts, and a robot cost line of 34,000 a year covering one cleaning unit plus its maintenance and power.
| Occupier | Net internal area (m2) | Share | Annual robot cost line 34,000 |
|---|---|---|---|
| Tenant A, floor 2 | 2,450 | 26.1 percent | 8,860 |
| Tenant B, floor 3 | 1,980 | 21.1 percent | 7,160 |
| Tenant C, floor 4 | 3,100 | 33.0 percent | 11,210 |
| Tenant D, ground retail | 640 | 6.8 percent | 2,310 |
| Common parts and core | 1,230 | 13.1 percent | 4,460 borne across the pool |
The method is transparent and cheap to administer, and its weakness is equally obvious: the cost follows floor area while the work follows use. Tenant C occupies a third of the building and might be a low-footfall back-office operation, while Tenant D occupies 6.8 percent of the area as a ground-floor retail unit generating most of the pedestrian traffic that a lobby machine spends its day negotiating. Area apportionment charges C the most and D the least, which is exactly backwards against benefit.
A well-advised tenant will make precisely that argument. The defence is consistency: if every other common-part cost in the building is apportioned by area, a single cost line apportioned by footfall is the anomaly. The counter-defence is that a robot is not a passive common part but an active service whose intensity varies with use, which is why many leases now permit a service-specific basis for a defined class of costs.
Method Two: Apportionment by Footfall or Occupancy
Footfall apportionment allocates the robot cost according to measured use. It is fairer on the argument above and considerably harder to administer, because it requires a defensible measure that tenants will accept.
The measure has to be a number both sides can audit. Three are commonly available:
- Door-count data from the access control system. Reliable, already recorded, and excludes visitors unless the visitor system records them too. It measures entries rather than presence, which overweights an occupier with many short visits.
- Occupied desk-hours from the space management system. Better for buildings with desk booking, because it captures duration rather than entries, but it depends on booking discipline and undercounts unbooked attendance.
- Delivered-parcel and dispatch volumes from reception logs. The most directly relevant measure if the machine's job is internal delivery, since it counts the actual service events rather than a proxy for them.
Worked on the same building with access-control door counts over a representative quarter:
| Occupier | Quarterly entries | Share | Annual robot cost line 34,000 | Change vs area method |
|---|---|---|---|---|
| Tenant A | 38,400 | 19.2 percent | 6,530 | minus 2,330 |
| Tenant B | 24,700 | 12.4 percent | 4,220 | minus 2,940 |
| Tenant C | 41,900 | 21.0 percent | 7,140 | minus 4,070 |
| Tenant D, retail | 94,600 | 47.4 percent | 16,110 | plus 13,800 |
Tenant D's charge more than doubles. That is the honest answer to the fairness question, and it is also the reason the method is politically difficult: the occupier who generates the most common-part use is the one who objects hardest, and in a retail context they may have a rateable-value argument of their own. Footfall apportionment is therefore usually applied to a subset of costs, agreed at lease grant, rather than imposed on a sitting tenant mid-term.
For the operational side of how a shared machine allocates its day between ground-floor retail traffic and upper-floor office use, the sequencing logic in multi-floor building deployment, and the wider pattern for smart office buildings and corporate campuses is the practical companion, because the schedule is what generates the benefit being apportioned.
Method Three: Benefit-Received Bands
The third method abandons computed fairness in favour of a negotiated schedule, and in practice it is the one that gets signed. The building is divided into benefit bands, each band carries a weighting, and the weights are set out in the lease schedule so that no annual recalculation or argument is needed.
| Band | Definition | Weighting | Effect on the share |
|---|---|---|---|
| Band 1 | Occupiers whose demised premises front the serviced common areas and receive direct service | 1.8 | Area share multiplied by 1.8 before normalisation |
| Band 2 | Occupiers on serviced floors with shared facilities but no frontage | 1.2 | Area share multiplied by 1.2 before normalisation |
| Band 3 | Occupiers with back-of-house or self-sufficient access | 0.6 | Area share multiplied by 0.6 before normalisation |
Weightings are applied to the area share, so a Band 1 occupier of 1,000 square metres pays three times what a Band 3 occupier of the same area pays for this cost class. The advantage is certainty: the schedule is fixed, the tenant can price the obligation at lease grant, and the landlord does not have to defend a new dataset every year. The disadvantage is that it stops responding, and an occupier that changes its use pattern has no route to a recalculation until the next rent review.
The three methods are not mutually exclusive. A common structure is area apportionment as the base, a single benefit-band adjustment for a defined class of active services, and a cap on the annual increase in that class so no tenant faces a surprise. That combination is what most institutional landlords end up with, and it is worth proposing at lease grant rather than at year two.
The Lease Wording That Makes the Recharge Enforceable
Whether any of this is recoverable depends on the service charge clause, and the difference between a recoverable and a non-recoverable cost is often one phrase. Four elements need to be present.
- A technology-neutral sweep-up. Clauses that enumerate cleaning, security and reception duties in fixed wording will not capture a machine unless the list is open-ended. The language to look for is a general provision covering equipment and services for the management and operation of the common parts, rather than a closed list of tasks.
- An equipment and plant provision. Many leases permit recovery of plant installed for the common parts but are silent on mobile equipment that is not fixed. A machine that lives in a store room and moves is not obviously plant, and a tenant's surveyor will notice.
- An apportionment mechanism for a defined cost class. If a service-specific basis is intended rather than pure area, the lease has to say so and identify the class. Adding a footfall basis to a lease that only contemplates area apportionment is the most common defect in these schedules.
- Disclosure of the underlying data. Where a cost is apportioned on a measured basis, the lease should require the landlord to make the measurement available on request. Without that the apportionment is not auditable, and a tenant's auditor will challenge it as unsubstantiated whether or not the number is right.
The practical consequence for a landlord planning a fleet is to read the lease before the specification. If the sweep-up is closed, the first building in the portfolio needs a variation, and a variation agreed as part of a lease event costs a fraction of one negotiated mid-term under pressure.
What Tenants Challenge, and How to Answer
Three challenges recur, and each has a defensible answer if the case was built properly.
The substitution argument. A tenant will point out that the machine replaced labour that was already being charged, so the service charge should fall rather than rise. The answer is to show the labour line falling in the same budget. A property that introduces a machine and leaves the cleaning contract unchanged has not passed on the saving, and the tenant is entitled to ask why. The corridor-cost arithmetic behind that comparison is in the cost model for common-area robotic service.
The benefit argument. An occupier on a self-sufficient floor with its own entrance and its own contractor receives no benefit from a lobby machine and will argue for exclusion. Band weighting is the mechanism that handles this, and a schedule that puts such an occupier in Band 3 answers the challenge before it is made.
The obsolescence argument. An occupier on a five-year term will object to paying for a machine depreciated over eight years with a residual value they cannot realise. Either depreciate over the shorter of asset life and average lease term, or use a subscription so the cost tracks the contract. This argument is usually the strongest one a tenant has, and it is the easiest to pre-empt.
Worked Case: One Machine, Four Tenants, Three Years
Assemble the decisions on the same 9,400 square metre building.
| Decision | Choice | Reason |
|---|---|---|
| Purchase structure | Subscription over 36 months, no capital | Average lease term 4.2 years, so the residual-value argument is avoided |
| Cost line | 34,000 a year inside the cleaning heading | Same heading as the labour it displaces, so the substitution argument can be answered line by line |
| Apportionment | Area base with a Band 1 to 3 weighting for active services | Consistent with other common-part costs, responsive to frontage benefit, fixed enough to price at lease grant |
| Cleaning contract adjustment | Contract hours reduced 18 percent at the next break | The saving has to appear in the pool or the recharge is indefensible |
| Data disclosure | Machine duty log shared quarterly | Gives occupiers a way to audit the apportionment without appointing a surveyor |
| Review trigger | Reapportion at each lease event | Cheaper than an annual recalculation and sufficient for a four-tenant schedule |
The item that makes this schedule work is the fourth row. A machine that appears in the service charge while the cleaning contract remains at full hours is a cost increase for occupiers and a saving for nobody, and any tenant surveyor will find it within an hour of the first budget meeting. Presenting the contract reduction alongside the new cost line converts an argument about a new charge into an argument about a smaller net increase, which is a much easier conversation.
Where the building also carries a security line, the same mechanism applies to a patrol machine, and how a patrol route interacts with cleaning hours is covered in patrol robot security rounds. The apportionment mechanics carry across unchanged.
What to Settle Before the Next Budget Cycle
- Read the service charge clause first. If the sweep-up is a closed list, plan a variation at the next lease event rather than a challenge mid-term. Do this before selecting a machine, because a building that cannot recover the cost should not be in the first wave.
- Decide the structure before the price. Subscription versus capital changes the apportionment argument more than the machine choice does, and the depreciation treatment is what occupiers challenge.
- Reduce the labour line in the same budget. A machine on top of unchanged contract hours is a cost increase, and the whole case depends on showing the net.
- Pick one measurement basis and disclose it. Measured apportionment without published data is worse than area apportionment, because it looks arbitrary and cannot be audited.
- Set the review trigger in the lease. Reapportioning at lease events is cheaper and more defensible than an annual recalculation nobody asked for.
- Budget the metering, not just the machine. Common-part power, dock circuits and connectivity all enter the same pool, and the electrical arithmetic in fleet charging infrastructure and power planning is the part facilities teams discover late.
In a multi-let building the robot is the easy part. The apportionment schedule is what decides whether the landlord can buy, and it is settled by documents written years before anyone thought a machine would be cleaning the lobby.
