7 Min. read time<\/p>\n
CSRD implementation in SMEs rarely fails due to lack of will\u2014almost always, it\u2019s the data model that collapses. Those treating the reporting obligation as an extension of the management report build a system that crumbles under scrutiny in year two. This article reveals where ERP bolt-ons fail on double materiality and which architectural decisions actually hold up.<\/strong><\/p>\n Key takeaways<\/p>\n Related<\/span>E-invoicing 2026: Mandatory rollout for SMEs<\/a> \/<\/span> Industrial heat transition: Chemical sector goes electric<\/a><\/p>\n Right now, many SME offices have a project plan titled “CSRD implementation” on their desks. Next to it sits the ERP system, a new sustainability module from the software provider, and a consulting offer promising “CSRD readiness in 90 days.” What\u2019s often missing is one sobering question: Will the data architecture we\u2019re building actually support the full reporting scope for the next five years?<\/p>\n The Corporate Sustainability Reporting Directive (CSRD) isn\u2019t just an expanded version of the management report. It\u2019s a standalone reporting regime with its own audit requirements and a data structure that shares little with traditional financial data. Underestimate this, and you\u2019ll pay the price in year two\u2014not with a fine, but with an audit finding that demands three months of rework.<\/p>\n This isn\u2019t another CSRD overview. It\u2019s an architectural critique. Why ERP bolt-ons fail on double materiality, which data models actually work, and where SMEs systematically underestimate the scope of their obligations.<\/p>\n Double materiality is the assessment concept underpinning the CSRD. It requires two perspectives for every sustainability topic: financial materiality (how the topic impacts the company) and impact materiality (how the company impacts people and the environment through the topic). A topic is considered material\u2014and therefore reportable\u2014if either perspective exceeds the defined threshold. The methodology must be documented and verifiable.<\/p>\n Source: EFRAG ESRS 1\/2, CSRD text.<\/p>\n<\/div>\n Since 2023, major ERP providers have added sustainability modules to their portfolios. Commercially, this makes sense\u2014and technically, they often deliver when it comes to quantitative environmental data: electricity consumption, diesel volumes, waste generation, business mileage. These data points already live in ERP systems, embedded in invoices, fleet records, or energy bills.<\/p>\n The problem starts where the ESRS demands qualitative and process-related information. Grievance mechanisms along the supply chain. Documented impacts on communities near production sites. Governance decisions on climate risks, complete with meeting minutes. These data points don\u2019t reside in ERP systems. They\u2019re scattered across HR platforms, legal departments, unstructured email archives\u2014or simply don\u2019t exist in any structured form.<\/p>\n An ERP bolt-on pretends sustainability reporting follows the same accounting logic as financial bookkeeping. That holds true for a fraction of data points. For the rest, the module offers sleek input masks where employees manually enter data they\u2019ve cobbled together in Excel. The audit trail remains thin because the actual evidence isn\u2019t in the system.<\/p>\n Double materiality lies at the heart of the CSRD. It requires documenting two perspectives for every potentially relevant topic: How does the issue impact the company (financial perspective)? And how does the company, through this issue, impact people and the environment (impact perspective)? Both assessments must be consistent, traceable, and audit-ready.<\/p>\n \nThe biggest risk isn\u2019t failing an audit. It\u2019s a data model that buckles under the second wave of additional ESRS fields\u2014triggering a follow-up project before the first one is even finished.\n<\/p><\/blockquote>\n Many projects treat the double materiality analysis as a one-off workshop. Two days with external consultants, a final document, job done. That\u2019s cutting it too short. The materiality analysis determines which ESRS data points ultimately need reporting\u2014and which don\u2019t. It sets the scope for the data model. If the analysis isn\u2019t structurally embedded in the system, it has to be rebuilt every year, and comparability is lost before the first report is even audited.<\/p>\n A robust data model maps the materiality matrix as its own entity. Topics, impact assessments, financial assessments, linked ESRS data points, evidence sources, responsible parties, review cycles. This isn\u2019t an Excel spreadsheet\u2014it\u2019s a relational structure. Ignore that, and by year three, you\u2019ll be explaining to your auditor why a topic\u2019s materiality changed between reports without anyone documenting the shift.<\/p>\n When talking to managing directors, it becomes clear that the effort required is consistently underestimated in three key areas. First, Scope 3 emissions\u2014those generated across the value chain. The figures come from supplier data, transport documents, and estimation models, yet they rarely align. Second, value chain information on social issues, such as working conditions at suppliers. Often, this data simply doesn\u2019t exist, and gathering it can take months. Third, governance documentation, particularly the link between management\u2019s strategic decisions and identified sustainability risks.<\/p>\n The blind spot is especially pronounced when it comes to Scope 3. A manufacturer with 200 employees typically sources from hundreds of suppliers, of which perhaps twenty are large enough to provide their own CSRD or comparable data. The rest report energy consumption on request\u2014figures that land in a single Excel cell, unchecked. Those who accept these numbers at face value build a reporting section that collapses under scrutiny the following year. Instead, those who clearly distinguish between measured, derived, and estimated emissions give auditors a traceable foundation\u2014even if the final figure looks less impressive.<\/p>\n The same applies to the social aspects of the European Sustainability Reporting Standards (ESRS). The standards require concrete statements on working conditions, discrimination complaints, and employee representation involvement. Companies with well-documented HR processes already have this material on hand. For others, it takes months to compile. This effort isn\u2019t just about data\u2014it\u2019s about the organization behind it.<\/p>\n The point isn\u2019t that SMEs must report all these data points\u2014most companies end up with no more than 300 material ones. The key is that the selection must stem from the materiality analysis, with every decision justifiable to auditors if challenged.<\/p>\n From projects that have become technically viable, three architectural principles emerge. They may sound unremarkable. Yet they make the difference between a system that can be expanded in year two\u2014or one that needs a complete rebuild.<\/p>\n What bolt-on approaches break<\/p>\n What a viable architecture supports<\/p>\n First: Evidence before value.<\/strong> Every reported data point gets an evidence anchor in the system\u2014a reference to the source document, the responsible person, and the date of last verification. Without this link, audit effort doubles the following year.<\/p>\n Second: Materiality matrix as a first-class entity.<\/strong> The matrix isn\u2019t an appendix to the report; it\u2019s the controlling structure behind it. Changes to the matrix require justification and versioning. Every reported data point ties back to a matrix entry.<\/p>\n Third: Separation between operational measurement systems and reporting layer.<\/strong> Raw data comes from diverse sources and undergoes operational changes. The reporting layer reads this data in fixed time windows, versions it, and preserves it even if source systems evolve. Mix raw and reporting layers, and traceability vanishes by year three.<\/p>\n The choice between an ERP module and a dedicated sustainability platform is often framed as a fundamental decision. In reality, it\u2019s far less dramatic. A bolt-on solution may suffice for companies whose reporting scope is strictly limited to quantitative environmental data and whose materiality analysis remains stable. That applies to only a minority of mid-sized firms.<\/p>\n What matters isn\u2019t the tool, but the sequence. Companies that pick software first and then adapt their data model to it inherit the vendor\u2019s limitations. Those that design their data model based on their own materiality first\u2014and then evaluate software against it\u2014end up with a leaner, better-fitting solution.<\/p>\n The February 2025 omnibus delay eased the time pressure for the second and third waves\u2014but it didn\u2019t make the content any easier. Companies that use this extra time to build a robust data model will be in a far stronger position in two years than those that simply bolt on a quick fix and hope it\u2019s enough.<\/p>\n Guidelines only\u2014status as of April 2026. Actual steps depend on company size, wave allocation, and auditor expectations.<\/p>\n<\/div>\n The most honest advice from companies already reporting in the first wave? Invest early in materiality analysis as a structural foundation\u2014not just workshop output. Define your evidence anchors before data collection begins. And keep reporting layers separate from operational measurement systems, even if it feels like overhead at first. By year three, that overhead will determine whether your audit takes three weeks or two quarters.<\/p>\n The Corporate Sustainability Reporting Directive (CSRD) isn\u2019t a compliance checkbox to tick in 2026 and forget by 2028. It\u2019s a new reporting regime that evolves every year. Get the architecture right now, and you\u2019ll avoid the costly rebuilds that others will face as early as year three.<\/p>\n Following the EU Commission\u2019s Omnibus package in February 2025, reporting deadlines for the second and third waves were extended. The exact timeline depends on employee count, balance sheet total, and revenue, as well as national implementation. Companies exceeding the thresholds should align the current transition periods with their auditor.<\/p>\n<\/details>\n For companies with a narrowly defined reporting scope and a stable materiality analysis, an ERP module may suffice. However, once qualitative social data, Scope 3 emissions, or consolidated data from multiple subsidiaries come into play, bolt-on solutions hit structural limits. The key isn\u2019t the tool itself, but whether its architecture can map the materiality matrix as a standalone entity.<\/p>\n<\/details>\n Three patterns keep recurring. First, the materiality analysis is treated as a one-off workshop rather than an ongoing structural foundation. Second, Scope 3 data is adopted without distinguishing between measurements, secondary data, and estimates. Third, there\u2019s no evidence anchor linking each reported data point to its source document, responsible party, and verification date.<\/p>\n<\/details>\n The ESRS framework includes around 1,000 data points across all standards. However, only those identified as material through your company\u2019s materiality analysis are mandatory. In typical mid-sized projects, the final report ends up covering between 150 and 300 data points. The selection process must be methodically documented for transparency.<\/p>\n<\/details>\n The Omnibus postponement has relaxed the application timeline for the second and third waves and announced simplifications for certain requirements. However, the core of the standards\u2014including double materiality and the audit obligation\u2014remains unchanged. Companies that use the extra time to build a robust data model will have a clear advantage in their second reporting year.<\/p>\n<\/details>\n Editor\u2019s Picks<\/p>\n\n
What is double materiality?<\/h2>\n
\n\n
\n \nPerspective<\/th>\n Question<\/th>\n Typical data source<\/th>\n Verifiability<\/th>\n<\/tr>\n<\/thead>\n \n Financial materiality<\/strong><\/td>\n How does the topic affect the company\u2014risks, opportunities, balance sheet?<\/td>\n Risk registers, controlling, business cases<\/td>\n Established, follows management report logic<\/td>\n<\/tr>\n \n Impact materiality<\/strong><\/td>\n How does the company affect people and the environment through the topic?<\/td>\n Audits, supplier assessments, stakeholder interviews<\/td>\n New for many companies; methodology must be documented<\/td>\n<\/tr>\n \n Combined assessment<\/strong><\/td>\n A topic is material if either perspective exceeds the threshold<\/td>\n Materiality matrix as working document<\/td>\n Thresholds must be self-defined and justified<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n Why ERP bolt-ons are the wrong answer to the wrong question<\/h2>\n
Double materiality is a data model, not a consulting exercise<\/h2>\n
Where SMEs Systematically Underestimate the Effort<\/h2>\n
Three data architecture decisions paying off now<\/h2>\n
\n
\n
Bolt-on or standalone platform? The more honest answer<\/h2>\n
When each architecture works<\/h3>\n
\n
\n
What now matters most<\/h2>\n
Frequently Asked Questions<\/h2>\n
When do mid-sized companies first need to report under CSRD?<\/strong><\/summary>\n
Is my ERP provider\u2019s sustainability module enough for CSRD compliance?<\/strong><\/summary>\n
What are the most common mistakes in CSRD implementation?<\/strong><\/summary>\n
How many ESRS data points does a mid-sized company actually need to report?<\/strong><\/summary>\n
How does the EU Omnibus initiative change CSRD obligations?<\/strong><\/summary>\n