Sustainability data for customers: the voluntary EU standard

The first step was a recommendation. On 30 July 2025 (European Commission) the Commission adopted Recommendation (EU) 2025/1710 on a voluntary sustainability reporting standard for small and medium-sized enterprises, known as VSME. It was developed by EFRAG, the Commission's technical advisory body for sustainability reporting. The Commission encouraged large companies and financial institutions to base their requests to SMEs on this standard as far as possible (European Commission). It was not an obligation, and that was exactly its weakness: anyone used to their own questionnaire could keep it.

The Omnibus I Directive of 24 February 2026 (Directive (EU) 2026/470) changed that. It required the Commission to adopt a voluntary standard as a delegated act, built on the 2025 recommendation and doubling as a cap on data requests. The Commission adopted Delegated Regulation (EU) 2026/1560 on 3 July 2026; it was published in the Official Journal on 21 September 2026 and has been in force since 24 September 2026 (Wirtschaftsprüferkammer, the German Chamber of Public Accountants). The legal text simply calls it the voluntary standard; in professional discussion it is VS, short for Voluntary Standard. With its entry into force the VSME recommendation is no longer considered legally effective (Delegated Regulation (EU) 2026/1560, recital 5). Anyone who has already reported under the VSME does not start over: the new standard keeps the structure of a basic module and a comprehensive module.

The framework is changing for the customers at the same time. The revised European Sustainability Reporting Standards, ESRS, enter into force on 10 November 2026 and must be applied at the latest to financial years beginning on or after 1 January 2027 (Wirtschaftsprüferkammer). If you supply a reporting company, you can therefore expect its questionnaires to be redesigned for financial year 2027. That is the right moment to organise your own data according to the voluntary standard.

The core for SMEs is Article 3 of the regulation: the voluntary standard sets the value chain cap. Reporting companies may not ask companies in their value chain that had an average of no more than 1,000 employees in the previous financial year for information beyond it for their sustainability reporting (Delegated Regulation (EU) 2026/1560, Art. 1). Annex II of the regulation lists the data points covered by the cap. Article 3 applies to financial years beginning on or after 1 January 2027 (Delegated Regulation (EU) 2026/1560, Art. 4).

The directive adds two rules to this right. Contract clauses that ask for more for the purpose of sustainability reporting are void, while the rest of the contract remains valid. And anyone asking for more must tell the protected company which items go beyond the standard and that it may refuse them (Directive (EU) 2026/470, Art. 2). These rules sit in a directive and take effect through national law; member states must transpose them by 19 March 2027 (Directive (EU) 2026/470, Art. 5).

Most values in the basic module already exist in the business, just not in one place. Electricity consumption is on the supplier's annual invoice; the standard explicitly refers to electricity as reflected in utility billings (Delegated Regulation (EU) 2026/1560, Annex I para. 32). Gas, heating oil and diesel appear in invoices and fuel receipts, and sub-meters show which site or hall uses how much. The catch lies in the units: the kilowatt hours on a gas invoice usually refer to the gross calorific value, whereas many emission factors refer to the net calorific value. Germany's Federal Office for Economic Affairs and Export Control converts natural gas from gross to net calorific value with a factor of 0.906 (BAFA) and states explicitly that its factors refer to the net calorific value.

Consumption turns into emissions by multiplying each type of energy by an emission factor. Scope 1 covers emissions from sources the company owns or controls, such as heating, furnaces and the vehicle fleet; scope 2 covers emissions from generating purchased energy, calculated location-based in the basic module (Delegated Regulation (EU) 2026/1560, Annex I para. 33). For the German electricity mix, the Federal Environment Agency estimates an average of 344 grams of CO2 per kilowatt hour for 2025 (Umweltbundesamt). In its funding programme, BAFA uses 0.201 tonnes of CO2 per megawatt hour for natural gas and 0.266 tonnes for light heating oil and diesel (BAFA). What matters is recording the factors used, with source and version, in the dataset and flagging any change of source. Otherwise emissions rise or fall from one year to the next without anything having changed in the business.

The worked example shows the logic, not a typical business. Gas is first converted to the net calorific value, diesel is translated into energy at 10 kilowatt hours (BAFA) per litre, and then each quantity is multiplied by its factor. If meter readings and invoices are already recorded in accounting, this needs no new software, just a fixed calculation rule in one place that is applied the same way every year.

Employee figures come from payroll or the HR system. The standard leaves the choice between headcount and full-time equivalents (Delegated Regulation (EU) 2026/1560, Annex I para. 40); what matters is to record the chosen method and the cut-off date so that next year's figure remains comparable. Contract type and gender are kept in the payroll master data anyway and can usually be exported as a list. Training hours, by contrast, are often not recorded completely anywhere. They are a typical case for a small new capture step, for example directly on the attendance record of a training session or safety briefing.

For work accidents, an existing obligation helps. The standard treats a work accident as recordable if it results in death or in incapacity for work of more than three days (Delegated Regulation (EU) 2026/1560, Appendix A). That matches the accident report to the statutory accident insurer in Germany: it is due if insured persons are injured so that they are unable to work for more than three days, and must be filed within three days of the employer learning of the accident (German Social Code Book VII, Section 193). If you file accident reports in an orderly way, you already have the figure for B9. The rate also needs a reference value, such as the hours worked from payroll or time recording.

Where large companies or banks require an annual update, the standard provides for the report to be prepared annually, for a period consistent with the financial statements (Delegated Regulation (EU) 2026/1560, Annex I para. 18). If data points have not changed, the business may refer to the previous year's value. From the second reporting year onwards, comparative figures for the previous year are included (Delegated Regulation (EU) 2026/1560, Annex I para. 14). This creates a fixed cycle that follows the financial statements the business prepares anyway. Much of what is prepared in December for the annual financial statements, such as accrued energy invoices and headcount at the cut-off date, also provides the basis here.

For a mid-sized business, three roles are usually enough: one person who owns the dataset and answers the questionnaires; the departments that supply their values, namely accounting, HR, facilities or fleet and occupational safety; and management, which approves and decides what is disclosed beyond the standard. Writing down ownership, sources and calculation once makes the cycle independent of individual people. That is process documentation on a small scale, and it also helps when the responsible person changes.

For most businesses the dataset is not a question of technology but of order: which figure applies, who supplies it, when is it ready. We therefore start with a short process analysis: which requests came in last year, which questions recur, where do the values sit today. This produces the mapping to the basic module and a list of gaps, which typically means training hours, hours worked for the accident rate or missing sub-meters.

We then build the dataset where your metrics are created anyway. Values from accounting, payroll and meter readings are collected through exports or interfaces, conversions and factors are stored in one place, and a report shows the status at the cut-off date together with the previous-year comparison. It is the same set-up as reports generated automatically, just with different fields. How such a metrics sheet comes about with us is described on the page metrics and reporting. The result is a dataset from which the next request can be answered without starting from zero again.