CO2 Monitoring in Buildings: Why Measured Data Beats Estimates

Measured CO₂ data from energy and fuel invoices show what a building actually emits. Flat-rate estimates, by contrast, rely on building types and standardized reference values. Once investment prioritization, ESG reporting obligations or proof of savings are on the line, measured, operational CO₂ figures are far more reliable than any blanket estimate. For a first overview, though, an estimate is often good enough.

For housing companies, this distinction stopped being academic a while ago. Since January 1, 2023, the Carbon Cost Allocation Act (CO₂-Kostenaufteilungsgesetz) has required landlords to determine actual, invoice-based CO₂ emissions per square meter of living space per year. If you are steering a portfolio, setting retrofit priorities or managing ESG reporting today, you can hardly get around solid numbers.

Why Does CO₂ Monitoring in Buildings Deliver More Reliable Figures Than an Estimate?

Measured CO₂ monitoring relies on real consumption data from meter readings and utility invoices, while estimates rely on standardized building-type models. The Carbon Cost Allocation Act turns this distinction into a legal requirement: a building's specific CO₂ value is calculated from the actual annual CO₂ quantity stated on the energy supplier's invoice, then split between landlord and tenant using a ten-tier model. For 1,000 square meters of living space and an annual quantity of 28,800 kilograms of CO₂, that works out to a specific value of 28.8 kilograms of CO₂ per square meter per year, a figure that comes straight off the invoice.

An evaluation commissioned by Germany's Federal Ministry of Transport, Building and Urban Affairs shows just how large that gap can be: among 94 energy certificates examined, only 29 percent of demand certificates fell within the permitted 5 percent tolerance of actual energy demand, compared with 66 percent of consumption certificates. The maximum deviation reached 108 percent for demand certificates and 26 percent for consumption certificates, as the analysis published in the IKZ trade journal shows. A smaller sample analysis of eleven buildings confirms the pattern: calculated final energy demand under DIN 4108 deviated from actual consumption by an average of 20.8 percent, reaching 56.5 percent in one case, and in 10 of the 11 buildings, actual consumption exceeded the calculated demand.

Demand vs. consumption: A demand certificate calculates a building's energy demand from its components, insulation values and standardized usage assumptions. A consumption certificate is based on actual billed consumption data from recent years. For operational decisions, the latter is by far the more reliable basis.

What Sources of Error Distort CO2 Estimates in the Existing Building Stock?

Several factors cause flat-rate estimates in the existing stock to miss the mark on a regular basis, and none of them can be captured by a type-based calculation alone: weather, heating behavior, operational faults and the individual impact of specific measures.

Weather is the most obvious factor. Weather adjustment of heating energy consumption uses official climate factors from Germany's national weather service, the Deutscher Wetterdienst, for more than 8,200 postal delivery districts. Because the previous method carried inaccuracies, the BBSR published a more precise calculation method in 2025 based on hourly grid data. Even with corrected figures, residual uncertainty remains: a milder or colder winter can mathematically adjust an actually measured saving of 25 percent down to 20 percent, or distort it the other way, depending on which climate factor is applied.

Residents' heating behavior weighs at least as heavily. As a rough rule of thumb from an older study initiated by ista and the Institut Wohnen und Umwelt (IWU), consumption in identical apartments in older buildings varies by more than 50 percent through user behavior alone, regardless of the building's physical condition. An estimate that only factors in building type and construction year cannot capture that range by design.

The effects of individual retrofit measures also differ noticeably from building to building. An evaluation of 108 multi-family buildings as part of the MOBASY study by the Institut Wohnen und Umwelt shows that physically calculated demand and measured consumption align fairly well on average across many buildings, but the deviation stays substantial in individual cases. That is exactly the problem for operational decisions: a portfolio average says little about whether a specific retrofit measure actually worked in one particular building. Operational faults, such as an incorrectly set heating curve or a night setback that stopped functioning, widen this spread further and can only be identified through real operating data.

When Do Housing Companies Need Granular CO₂ Data for Portfolio Decisions?

Granular, building-level CO₂ data becomes essential whenever capital, reporting obligations or investment sequencing are on the line. Three situations come up especially often in the housing industry: ESG reporting duties, financing or stranding-risk assessments, and proof that a specific measure actually worked.

Under CSRD and ESRS E1, affected real estate companies must disclose greenhouse gas emissions by Scope 1 and 2, and later Scope 3, on a building-by-building basis. The EU Commission's Omnibus reform package raised the threshold in 2025 to companies with more than 1,000 employees and either over €50 million in revenue or more than €25 million in total assets, cutting the number of companies subject to reporting by roughly 80 percent. For the companies that still have to report, often the larger ones, the core requirement barely changes: a rough type-based value is not enough for building-level disclosure. How this data perspective fits into a concrete decarbonization strategy is explored further in this interview on data for decarbonization.

On the financing side, the Carbon Risk Real Estate Monitor has become the de facto standard. CRREM provides more than 1,000 science-based decarbonization pathways in kilograms of CO₂ per square meter per year for more than 40 countries, and uses them to calculate the point at which a building falls off its pathway. Banks and institutional investors apply this assessment voluntarily, though in practice it has become binding for financing decisions. A stranding date is difficult to derive credibly from a building-type estimate, because it depends on the actual emissions trajectory of one specific building.

Good to know: CRREM pathways are voluntary, but banks and institutional investors increasingly use them as a screening criterion for financing. Without building-level consumption or emissions data, a position on the pathway is hard to substantiate.

Proving that a retrofit really worked demands the same depth of data. GdW, the umbrella association for German housing companies, illustrates how demanding even a basic data foundation can be: only through a dedicated 2025 special survey was it able to establish a reliable, consistent assignment of energy efficiency classes for more than 90 percent of all units across its roughly 3,000 member companies. If even a sector-wide classification only comes together through a targeted survey, it becomes clear why a single retrofit measure is hard to prove without building-level before-and-after data, especially with the retrofit rate in Germany's residential building stock having dropped to a new low of 0.67 percent in 2025.

When Is a Simplified CO₂ Estimate Still Enough to Get Started?

For an initial portfolio overview, or as a starting point ahead of a targeted investment, a simplified estimate remains a legitimate tool, especially where the technical infrastructure for granular measurement is still missing. As of December 31, 2025, only around 5.5 percent of Germany's nearly 56.5 million electricity meters were genuine smart metering systems with a smart meter gateway, even though the law calls for a 95 percent rate by 2032. For many buildings in the existing stock, seamless, near-real-time measurement is simply not available everywhere yet.

The Heating Costs Ordinance (Heizkostenverordnung) sets its own timeline here: by the end of 2026, all heat and hot water meters as well as heat cost allocators in multi-family buildings must be remotely readable, and from 2027, monthly remote reading including consumption information for users becomes mandatory. Until these deadlines are met, a type-based estimate is often the only basis available for establishing any kind of order within a portfolio, for example to decide which buildings deserve a closer look first. What matters is treating such an estimate as a preliminary starting point that still needs verification before it can carry an investment decision.

How Do You Move from CO₂ Estimates to Reliable Monitoring?

The path from estimate to reliable monitoring follows a fixed sequence, starting with the metering landscape, moving through prioritization, and ending with operational use of the data.

The first step is worth a sober look at your own metering landscape: which buildings already have remotely readable heat and hot water meters, and where gaps to the HKVO deadlines remain. The second step is prioritization: buildings with CSRD relevance, CRREM stranding risk or a particularly high CO₂ cost share deserve granular capture first, while the rest of the portfolio can keep working with an estimate for now. The third step covers the structured provision of this data, for example through a structured user portal for energy and consumption data, so the figures reach the property managers and technicians who actually have to work with them.

The fourth and decisive step is translating measured values into action. A CO₂ figure on its own does not save any emissions. It shows, above all, where a closer look pays off. Practical examples from the housing industry show how fast this can happen: at a housing cooperative and a municipal housing company, continuous digital heating monitoring enabled heating cost savings of around 20 percent within a few weeks, because misconfigurations and night setbacks that had stopped working could be identified from real operating data rather than blanket assumptions. This example comes from a single case, and the sample size is too small to represent a portfolio-wide average, but it shows how operational transparency, delivered for instance through the corresponding function in KUGU VIS, turns a CO₂ figure into a concrete operational decision. What this transition from measurement to digital heating optimization looks like technically in day-to-day operations is illustrated by a case example from a multi-family building with commercial units.

Reliable CO₂ Data as the Basis for Retrofit Sequencing and Capital Decisions

Honestly, measuring versus estimating is not really the question. The real question is: how precise does the number need to be for this particular decision? A rough portfolio overview can tolerate an estimate. For an ESG disclosure, a financing review or proof that a retrofit worked, it no longer holds up.

With the retrofit rate falling and the carbon price rising, getting this call wrong becomes more expensive. Allocating scarce retrofit budgets based on an imprecise type-based estimate risks leaving exactly the buildings with the highest stranding risk waiting too long for their turn. The pragmatic next step is not a complete overhaul of data management, but a targeted prioritization: capture the buildings with the greatest regulatory or economic leverage granularly first, run the rest of the portfolio on an estimate for now, and close the gap step by step.

Frequently Asked Questions About CO₂ Monitoring in Buildings

How Often Must Landlords Determine Their Buildings' CO₂ Emissions Under the Carbon Cost Allocation Act?

Annually, as part of the heating cost statement. Since January 1, 2023, the Carbon Cost Allocation Act has required an actual, invoice-based figure in kilograms of CO₂ per square meter of living space per year, which forms the basis for cost allocation under the ten-tier model.

At What Point Do CO₂ Data Become Relevant for CSRD Reporting Obligations?

They become relevant for companies that, under the EU Commission's Omnibus reform package, have more than 1,000 employees and either over €50 million in revenue or more than €25 million in total assets. For these companies, ESRS E1 requires building-level emissions data by Scope 1 and 2, and later Scope 3 as well.

How Accurate Is a Weather-Adjusted Consumption Measurement Compared to Actual Consumption?

Considerably more accurate than a pure demand estimate, though not error-free. The choice of climate factor applied can mathematically shift a real saving of 25 percent to 20 percent, or the other way around, which is why the BBSR published a more precise, hour-based calculation method in 2025.

Is a Consumption Certificate Enough to Make Retrofit Decisions?

Yes, for a first assessment, but usually not on its own for a concrete investment decision. According to a BMVBS evaluation, consumption certificates come noticeably closer to actual energy demand than demand certificates, but they do not show which specific measure produced which effect.

What Happens If Remote Reading Is Not Implemented by 2026?

Under the Heating Costs Ordinance, heat and hot water meters as well as heat cost allocators in multi-family buildings must be remotely readable by the end of 2026, and monthly remote reading with consumption information becomes mandatory from 2027. Until implementation is complete, housing companies often remain dependent on annual readings or estimated figures.