The landlord signs the contract; the tenant runs the building

In most of the offices on our platform, the landlord controls the central plant and the tenants control almost everything else. The landlord's contractor sets the boilers, heat pumps and air-handling units; the tenants decide when their floors are occupied, how many screens and kitchens run, whether the meeting rooms are cooled at midnight and how often someone props open a stair door. In the multi-let offices we have measured most closely, between 55 and 70 per cent of the electricity used passes through tenants' own meters, and the heating and cooling the landlord supplies responds largely to their hours and habits.

Yet the software market for buildings is organised almost entirely around the landlord. The landlord pays, the landlord's sustainability team writes the specification and the landlord's asset manager attends the demonstration. Tenants appear in the requirements as a stakeholder group, usually under the heading of engagement, and in practice they receive a branded portal and an occasional newsletter. I have built products that way myself. I now think it is the main reason building-performance software produces better reports than it produces buildings.

Who controls the consumption

The argument is simple once the numbers are laid out. A landlord facing the Dutch requirement for most offices to hold an energy label of at least C, or the minimum performance standards that the recast EU directive will bring for the worst-performing non-residential buildings, needs to improve the building's fabric and installations. That is the landlord's work, and our software helps to plan it. But a landlord who also reports actual consumption to investors and lenders, as most of our customers now do, is reporting a number produced largely by other people.

Consider an office of 12,000 square metres let to six tenants. The landlord can replace the heat pumps, add insulation and tune the building automation, and our fault detection will find schedules left in manual and valves that never close. In buildings where we have done all of that, consumption per square metre has fallen by about 11 per cent. The remaining reductions depend on tenants: switching off floors that stand empty on Fridays, setting server rooms properly, letting the building's controls work instead of overriding them. If the software never reaches the people who make those decisions, the landlord's report stays where it is, however good its charts.

There is a commercial reason as well. Commercial leases in the Netherlands, Belgium and Germany increasingly include clauses in which landlord and tenant agree to share consumption data and cooperate on improvements, often called green lease clauses. A clause is an intention. It becomes a practice only when both parties can see the same figures and have a reason to act on them. In our experience the tenant's facilities manager is the person who turns that intention into practice, and she is not the person most building software was designed for.

Three things a tenant will actually use

When we redesigned our occupier product in 2022, we began by asking tenants' facilities managers and employees what they would do with it. The answers were modest and consistent. They wanted to tell the building that a room was too warm or too cold, and to see that someone had responded. They wanted to report a fault without phoning a reception desk. And the facilities managers wanted to see their own consumption, compared with similar floors, in a form they could take to their own management.

That became the whole product: a comfort report, a fault report and a consumption view. We removed an events calendar, a community feed and a marketplace for local services that two large landlords had asked for, because almost nobody used them and each competed for attention with the three functions that mattered. Every page meets WCAG 2.1 level AA, and a tenant can use the comfort and fault reports from a code beside the lift without installing anything.

The comfort report turned out to be the most valuable source of data in the building. When forty people on a floor report that it is too warm on Tuesday afternoons, and the building automation shows that the cooling valve on that floor has reached full position, the fault is found in a day. Before, it would have surfaced as a complaint to the asset manager months later. Where the occupier product is active, the median time from a comfort report to an identified cause is now under three working days.

The consumption view changed behaviour more slowly, but more durably. Facilities managers who could compare their floors with others in the same building began asking the questions we had wanted landlords to ask: why the floor draws power at night, why the meeting rooms are cooled when empty. Several tenants have since requested changes to the building's schedules that the landlord's contractor would never have proposed.

What the landlord gives up

Designing for the tenant costs the landlord something, and a product leader should say what. The first cost is control of the message. A tenant who can see consumption can also see when the landlord's plant performs poorly, and some asset managers are uncomfortable with that. We built the consumption view so that tenants see their own figures and anonymised comparisons, not the landlord's whole-building data, but the discomfort remains, and two customers have declined to switch the view on.

The second cost is features the buyer wanted. The landlords who asked for a community feed were not wrong to want their buildings to feel like places people belong. They were wrong to think a software feed would achieve it, and we lost one tender in 2023 to a competitor that offered exactly the broad tenant app they described. I would make the same decision again, but I do not pretend it was free.

The third cost is time. A product for two audiences takes longer to design and longer to sell, because the landlord must persuade tenants to use it. We now include a tenant onboarding plan in every contract and judge our success partly by the share of tenants still active a year after installation, which in 2026 was 46 per cent across the buildings where the occupier product runs.

For a landlord choosing building software, I would suggest three questions in the demonstration. Who in the tenant's organisation will use this, and what will they do with it on an ordinary Tuesday? What will the tenant see that we cannot show them today? And what share of tenants are still using it a year after installation? A vendor that cannot answer has built a product for the person who signs, not for the people who run the building.