Sustainable retail operations

What this paper claims, and what it does not

Retail assets are among the most operationally intensive commercial buildings in Europe. The harder question is what better operation is worth to the owner, and this paper answers it in two separate parts.

The cost channel is directly measurable. Consumption falls, energy cost falls, and a defined part of that reduction reaches the owner's net operating income without anyone's agreement.

The customer channel is still to be tested. Indoor quality may change customer behaviour, and turnover rent may follow above the contractual breakpoint, but that has to be demonstrated for the specific asset.

The two are kept apart because one does not prove the other. Only the cost channel is settled. Footfall is tracked separately: catchment, access, tenant mix, events, marketing and public realm are major drivers.

What is coming anyway

Most of this paper is about what better operation is worth. Some obligations arrive without anyone deciding on them, and the data foundation they assume is the same one the commercial case runs on.

The EPBD building-automation obligation applies above a 290 kW technical-system threshold, falling to 70 kW at the end of 2029, where technically and economically feasible, and requires continuous monitoring and logging, benchmarking with loss detection, fault detection and interoperability. Under the EED, an energy audit above 10 TJ and a management system above 85 TJ attach to the enterprise on a three-year average, not to the individual building; a 50,000 m² centre accounts for roughly 54 TJ, so a portfolio reaches the threshold quickly. ISO 50001 certification discharges the audit obligation. Minimum energy performance standards follow national-stock trajectories, 16% by 2030 and 26% by 2033, which are not reduction targets for each individual building.

Alongside the two dated obligations sit market events in no fixed order: a lease renewal where the tenant compares total occupancy cost and energy performance and has alternatives, a refinancing where the lender looks past static certificates for evidence of actual performance, and a revaluation or sale where the valuer and buyer assess income resilience, transition risk and how well the building is run.

The operating cost channel

A centre at 06:00 does not need what it needs at 14:00, and a Monday morning does not behave like a Saturday afternoon. The gap between actual demand and building operation is where the potential sits: operating schedules that have drifted from actual trading hours, ventilation running to design rather than demand with no CO₂-based modulation, zones heating and cooling against each other, plant, common-area and car-park loads running while the centre is closed, and technical faults that persist for months because no system reports them.

Most of these are corrected by operating existing equipment differently, not by replacing it. Where a measure needs capital, this paper treats it as capital expenditure, not as an operating saving.

The worked example. A 50,000 m² shopping centre, illustrative throughout: 220 kWh/m² electricity and 80 kWh/m² thermal intensity, 15,000 MWh or 54 TJ of total site consumption, €0.18 per kWh electricity and €0.07 per kWh thermal, €2,260,000 of total site energy cost or €45.20 per m², a 45% landlord-metered share, a 10% structurally irrecoverable share, 8% vacancy, a 25% illustrative reduction on the landlord boundary and a 7.0% illustrative capitalisation yield.

A 25% reduction on the landlord boundary is €254,000 a year, but it does not arrive as one number: €25,000 is P1 irrecoverable, €18,000 is the P2 vacancy share and €211,000 is the P2 tenant share. Only €43,000 per year reaches the owner's NOI without requiring tenant consent, which at 7% supports roughly €625,000 of value sensitivity. P3, behind the tenant meter, is not addressed in this case.

From cost effect to NOI

Only two things enter NOI: owner-borne cost that falls, and a contractually secured recurring payment that a valuer will recognise as sustainable net operating income. A lower tenant service charge is neither. It improves the retailer's occupancy economics and supports retention and lettability, but it is not landlord NOI, and presenting it as such fails the first time a valuer tests it. Retention and lettability are often worth more than the NOI line; they are simply a different claim.

Where a green lease or performance clause allocates part of the verified tenant benefit back to the landlord, the picture changes. At a 0% share the owner sees €43,000 of NOI and roughly €625,000 of value at 7%. At 30%, negotiated at renewal, €63,000 additional NOI gives €107,000 in total and roughly €1,527,000. At 50%, under contracted shared savings, €105,000 additional gives €149,000 and roughly €2,128,000.

None of these allocations is a market standard. In European retail, negotiated sharing of this kind is the exception rather than the rule and is most readily agreed at renewal. The retention and lettability case holds in all three, including where no sharing is ever agreed.

The scale of the numbers. At €180 per m² average rent the centre produces roughly €9.0 million of gross rental income. That is 206 times the direct participation of €43,000, and roughly 35 times even the full €254,000 saving. The large number is the rent roll, and its durability is the reason to run the building better.

Turnover rent. Base rent is flat until the breakpoint, so only turnover above it produces additional rent. On illustrative inputs of €150m centre turnover, 60% turnover-rent penetration, 50% of those sales already above breakpoint, a 2% sales uplift and a 3% blended participation rate, €900,000 of incremental above-breakpoint sales gives €27,000 a year of additional rent and roughly €386,000 of value sensitivity at 7%. The 2% uplift is an assumed input, not a workcloud24 result. Both lease penetration and breakpoint status must be known; applying the participation rate to the whole sales uplift will not withstand a lease review.

Value sensitivity

Two different mechanisms are in play and they are never summed. The contractual NOI effect combines owner-borne cost reduction plus any secured recurring payment, capitalised at the applicable rate, giving roughly €625,000 to €2.13m from €43,000 to €149,000 of NOI at 7%. The market risk effect covers income resilience, lower transition risk and better evidence at refinancing and sale, and shows up in the yield if at all. The yield effect is a market observation and is not quantified here. Even the first is a sensitivity, not a valuation outcome: a valuer will discount it or capitalise it over a shorter horizon.

The customer channel

Retail is different because the building user is also the customer. The commercial question is not whether comfort matters, but how much measurable operating quality contributes in this specific asset.

The building-led chain runs from indoor environmental quality (thermal comfort, air quality, lighting, fewer disruptions) to customer behaviour (dwell time, conversion, basket size) to commercial outcome (tenant sales) to lease outcome (turnover rent, only where sales cross the contractual breakpoint). Footfall is a separate KPI, heavily influenced by catchment, access, tenant mix, events, marketing and public realm; measure it, but do not treat HVAC or a green label as the sole cause.

Attribution is tested by putting indoor environmental quality, footfall, dwell time, tenant sales and lease data on one time axis, controlling for weather, day type, season, tenant mix, promotions, events and works, and then testing whether changes in operating quality explain customer outcomes after those controls. External evidence supports a relationship between indoor environmental quality, shopping experience and retail performance. It does not establish the incremental effect of building operation in a specific asset.

The operational steering layer

workcloud24 connects existing building systems, metering, sensors and operational data into a continuous loop. The objective is not another one-time energy audit. The loop measures where, when and why energy is used, analyses loads, schedules and faults, optimises to actual demand, verifies against an agreed baseline, shares verified benefits and translates the result into data for reporting. Each step produces the evidence the next one needs, and optimisation runs inside a comfort envelope agreed before anything is changed.

Ten parallel data sources feed one model: building management systems, smart meters and IoT, HVAC and ventilation, lighting, weather data, energy data, PV, battery and EV where installed, trading and utilisation, footfall and dwell time where legally permitted, and tenant sales and lease data including turnover-rent clauses, breakpoints and participation rates.

Baseline and verification

Savings claims are only contractible if the measurement is. The baseline and its adjustments follow IPMVP principles: an agreed baseline period fixed before optimisation begins with the quality of the underlying data documented, routine adjustment for independent variables that move with normal operation, and non-routine adjustment for changes the building did not control.

In retail the energy model must control for weather, trading hours, occupancy, footfall, closures and refits and equipment changes, while the customer model must control for day type, season, tenant mix, promotions, events and works. Data gaps and non-routine adjustments are written down rather than quietly averaged away.

Retail-specific levers

Some levers exist in retail and not in office or logistics, though the relevance of each depends heavily on the asset. In food-anchored centres the anchor's refrigeration is the largest single load and the largest heat-rejection source at the same time. Regulation (EU) 2024/573 narrows refrigerant availability on a fixed timetable, forcing plant intervention. Roof and car-park surfaces offer large generation area with a demand profile peaking in daylight trading hours. Load shifting and peak management apply, with peak-demand management usually the larger and more reliable of the two effects; flexibility-market access depends on grid connection and aggregator terms and is an integration opportunity, not a revenue line.

Most refrigeration load sits behind the tenant meter, so it requires the tenant's participation and sub-metering. Roof and parking structures the landlord already owns can be contracted without amending the lease.

Carbon cost exposure. ETS2 prices fossil fuel used in buildings from 2028, postponed by one year. It does not cover purchased electricity, which is already priced through ETS1. For the example asset, thermal consumption of 4,000 MWh corresponds to roughly 810 tonnes of CO₂ a year, giving an annual exposure of €36,000 to €48,000 at a carbon price of €45 to €60 per tonne. Against €2.26 million of asset energy cost that is 1.6 to 2.1%, though it adds 13 to 17% to the fuel component alone.

Asset typology

The same levers do not apply everywhere. A shopping centre's cost boundary is P2, common areas billed through the service charge, its main lever is mall HVAC, ventilation and car-park lighting, and it has little leverage behind the tenant meters. A retail park's boundary is P3, unit-metered with minimal landlord plant, its lever is external lighting, roof generation and EV charging, and with no conditioned mall the service-charge argument is weak. A food-anchored centre mixes P2 and P3, its levers are refrigeration, waste-heat recovery and ventilation, and the landlord controls only part of the largest load.

Operations, communication and the model

Comfort, roles, availability and data rights are settled before optimisation starts, not discovered afterwards. Optimisation runs inside a comfort envelope agreed in advance to EN 16798-1 categories, and during trading hours comfort takes priority over consumption. Roles are defined before any change: who operates, who may change a setpoint and how a change is rolled back. Retail trades seven days a week, so a Saturday HVAC failure is a revenue event, not a maintenance ticket, and overrides stay manual. Data rights are contractual, not technical: ownership, permitted use, aggregation, retention and access are agreed before datasets are joined.

Footfall and dwell-time measurement requires a GDPR assessment, and pseudonymised data can remain personal data. Tenant sales and lease terms require contractual data rights; without them, customer-to-turnover-rent attribution cannot be tested.

Communicating the result. Directive (EU) 2024/825 restricts generic environmental claims to consumers from 27 September 2026 and requires them to be substantiated. A customer-facing display in a centre carries the same substantiation duty as any consumer claim, and every claim must state the asset boundary, metric, energy carrier, baseline period, reporting period and methodology, and whether the figure is measured or modelled. An owner with verified operating data can make the claim after September 2026; one without it cannot.

How it starts

Seven steps, on one asset. The pilot produces the internal evidence a portfolio decision needs, on the owner's own building rather than someone else's case study: an initial assessment from bills, meters and BMS, a building performance report, an agreed verification plan, optimisation of what already exists, verification and reporting per boundary, and only then an assessment of technology against set KPIs.

Document the position before, optimise, measure after. Sizing capital against an unoptimised load profile risks over-investment.