Water at 2.6: When a Group Average Hides a Material Impact
A European packaging manufacturer with 3,200 employees and 11 plants is preparing its first CSRD-aligned sustainability statement. Its double materiality assessment screened 24 topics, scored impact and financial materiality from 1 to 5 across two internal workshops, and applied a combined threshold of 3.5. Water withdrawals scored 2.6, because the scoring used group-wide abstraction volumes rather than local basin conditions. During a customer audit, a beverage client whose own plant sits in a catchment classified as high water-stress asked why water had been excluded from the material topics. Separately, the sales director has asked that "chemicals of concern in food-contact packaging" be dropped from the list, arguing that customers will read it as an admission of a problem. The sustainability lead has three weeks before the statement goes to the audit committee, no documented rationale for either position, and a first-year methodology that the external assurance provider has not yet seen.
1. Should water be treated as material? Set out how you would decide and what you would change.
Read the worked answer
Where the analysis went wrong. Impact materiality is assessed where the impact actually occurs, not as a group average. A score of 2.6 computed from total abstraction volumes hides the fact that two of the 11 plants sit in basins classified as high stress. A low group average can coexist with a material local impact, and severity (scale, scope, irremediability) should be scored per site.
Decision rule. A topic is material if it is material on either dimension. The 3.5 combined threshold is an internal screening tool, not a safe harbour. Re-score water at plant level using a published water-stress dataset: if even one site crosses the threshold on the impact dimension, water is material and must be reported.
What I would change. (1) Split the water datapoint by basin and by site. (2) Document the scoring method, the water-stress data source, and why the first pass missed the local context. (3) Report the impacts, risks and opportunities, any target, and the mitigation actions. (4) Keep the group-level figure only as context, never as the basis for exclusion. If water is genuinely not material after re-scoring, that conclusion must be evidenced and documented, not asserted.
2. How should the sustainability lead handle the sales director's request to remove the chemicals-of-concern topic?
Read the worked answer
Materiality is an evidence question, not a communications question. The request has to be tested against the actual and potential impacts, not against how a customer might read them.
- Test the claim. Is "chemicals of concern in food-contact packaging" associated with real or potential negative impacts - health effects, recyclability constraints, food-contact restriction exposure? If yes, it is likely material on the impact dimension regardless of commercial discomfort.
- Document the request. Record the date, the substance, who asked, the reasoning given, and your assessment. This protects both the reporter and the process when the auditor, a regulator or an NGO asks why a topic disappeared between drafts.
- Escalate. Take it to the audit committee as a governance matter. A request to remove a topic for reputational reasons is precisely what the committee exists to oversee.
- Reframe, do not suppress. The legitimate concern is narrative risk, and that is addressed with precise, proportionate wording - what the substance is, what substitution work is under way, by when - not with omission. Omitting a material topic from a management report is a misstatement, and if the statement sits inside the management report it carries legal consequences for the directors who sign it.
3. Was the design of the materiality threshold itself part of the problem? What would you recommend for year two?
Read the worked answer
A single combined threshold invites exactly this failure: a 4.8 impact score averaged with a 2.2 financial score can land below 3.5 and vanish from the report.
- Separate thresholds for impact materiality and financial materiality, so a topic can qualify on one dimension alone.
- Site or geography as the unit of analysis for location-specific impacts (water, biodiversity, local air emissions), aggregated only after scoring.
- Two-dimensional visualisation of results for the audit committee, rather than one ranked list that conceals the shape of the data.
- Sensitivity check: re-run the scoring under plausible alternative weightings and stakeholder inputs. If the material topic list changes, the methodology is too fragile to rely on.
- Calibration and continuity: record thresholds, weights, scorers and data sources, and explain year-on-year changes. Changing a threshold after seeing the results is acceptable only if the reason is documented and applied consistently.
- Traceable stakeholder input: a customer naming a real catchment is evidence; an undifferentiated workshop ranking is not.