Sovereignty and portability – the system requirement that matters for the future of your energy data

A requirements specification for energy management written in 2022 could be very thorough. Report templates, permission levels, integration with the property management system, requirements for how anomalies should be displayed. What it did not contain was a line about the data being machine-readable by an AI agent, because nobody was asking for that then. Today, it’s one of the most common questions asked when evaluating a tech stack. But the system procured in 2022 is still running.

A specification describes the present, the system has to last far longer

Public procurements of energy management are evaluated on functional requirements: a list of things the system should be able to do, scored and weighed against price. The requirements are written from how the organisation works at the time of the procurement, which is the only material there is to work from. It is the timescale that grates. Under EU procurement rules, a framework agreement in the classic public sector may as a general rule run for no more than four years, and longer only in exceptional cases that can be justified. A contract has no statutory time limit, as long as the term is proportionate. The meter structure and the history outlive both. The data you collect under this agreement will need to be used in at least one more system, probably several.

Anything that is not in the evaluation matrix carries no weight. A supplier who makes it easy to move data gets no points for it, and a supplier who makes it difficult loses none. But that does not mean it costs you nothing.

The cost of changing systems never appears in the tender

When the time comes to change systems, the history and the meter structure stay with the supplier who collected them. We regularly talk to technical managers and system owners in the middle of a migration who find that the export exists on paper, but takes months in practice.

Ownership of the data is usually in the contract already. Control over it rarely is. The difference shows up the day you want all of it out: can you do it yourselves, when you want, in a format someone else can read, or does the job have to be ordered from the party you are about to leave?

What has to be movable is the whole context. The meter point structure, the tags and the link between meter and building are what make the values intelligible in a new system. With continuous access to all of it throughout the contract term, you can migrate at a pace you set yourselves, test the data in a new system before the old one is terminated, and bring the years of history with you into the next procurement. If the only route to your data is a request to the supplier, leaving becomes a negotiation, and that negotiation happens at the point where you have the least leverage…

Nobody knew in 2022 what the data would be used for in 2026

Generative AI is the clearest example of why long lists of functions age. A few years ago the question was barely on the agenda at property organisations. Now the conversations are about whether you can put questions directly to the energy data, whether an agent can retrieve consumption for an entire portfolio without someone exporting a file first, and what needs to be validated in the data for the answers to be reliable.

The requirement that would have helped could have been written back in 2022, without knowing anything about what was coming: that all data must be retrievable by machine, in a documented format, with the history intact. A requirement like that does not age, because it describes access to the data and leaves the use open. The same goes for what comes after AI agents, whatever that turns out to be. You do not need to know exactly what will happen in order to write the requirement.

Five requirements to bring into your next specification

The requirements below are about your data, which is why they work whichever supplier you end up choosing. In Sweden, the Agency for Digital Government and the National Agency for Public Procurement have produced recommendations for procuring data together with several other bodies, including ready-made requirement texts, and several of the points below sit close to them.

1. Ownership of data and metadata. “The contracting authority owns all collected metering data and associated metadata, including meter point structure, tags and areas of responsibility. The supplier’s right of use is limited to delivering the contracted service.” Metadata takes years to build up and is expensive to recreate.
2. Export in an open format, without a separate order. “All data must be exportable in an open, machine-readable format, for example CSV or JSON, by the contracting authority itself and without consultancy work.” Otherwise every migration becomes an assignment priced by the party you are about to leave.
3. Access through a documented API throughout the contract term. “Throughout the contract term, the contracting authority must have access to all of its data through a documented API. The documentation must be sufficient for a third party to load the data.” This is the requirement with the longest shelf life, because it applies just as much to a BI tool, a sustainability platform or an AI agent.
4. The history follows. “The export must cover the full history registered in the system, regardless of which contract period the values were collected in.” Without several years of history there is no way to compare before and after a measure, and energy performance certificates and sustainability reporting are built on series over time.
5. Exit terms written into the contract from the start. “On termination of the contract, a complete export of data, metadata and structure must be delivered within a stated number of working days, at a cost stated in the tender.” The price of leaving and taking your data with you should be known while you are still comparing tenders.

Five lines in a specification cost nothing to write. They move the assessment from what the system puts on show today to what you can do with your data when the circumstances change.

And as long as collection and energy management sit in the same system, the same supplier owns both the data you want to move and the tool you are moving away from. There is a way around that.

Let the collection continue when the energy management system changes

Metry is a dedicated collection system. We collect, quality assure and structure the energy data, and leave energy management, sustainability analysis and tenant billing to those who are experts at it. You reach the data through an open API that is included for all customers, and through ready-made integrations to most of the larger energy management systems. Your ownership of your data is written into our terms and conditions, and the structure you build comes with you. So a system change becomes a change of analysis tool, while the collection carries on as usual.

Are you procuring a new system for your energy and sustainability departments? Get in touch and we will go through how to future proof your data collection.