CHANNELS    |    Behavior Change      Leadership      Products & Design      Supply Chain      Marketing & Comms      New Metrics    |    MORE


News & Views

Are Materiality Matrices Really Material?

While it is common practice now for corporate sustainability reports to include materiality matrices, whether or not they actually serve their purpose is debatable.  Indeed, we (the Center for Sustainable Organizations) don’t think they do, and have some suggestions for how to improve them.

The relevance of materiality matrices to sustainability measurement and reporting was perhaps first established by the Global Reporting Initiative (GRI). GRI, for its part, defines materiality as follows (G3.1, p. 8):

The information in a report should cover topics and Indicators that:

  • reflect the organization’s significant economic, environmental, and social impacts
  • or that would substantively influence the assessments and decisions of stakeholders.

At the very least, then, behaviors that result in significant economic, environmental and social impacts are material, or at least potentially so. It is important to add, we think, that the reason why such impacts are material is because they affect the interests and well-being of others (i.e., stakeholders).

This leads to GRI’s second dimension of materiality.  As presently worded, the second condition is rather vague — “assessments and decisions” about what, one might ask?  More important is the fact that materiality is first and foremost about the first dimension, and that any impacts that are deemed material should, by that standard alone, be of automatic interest to shareholders; which is what GRI’s second dimension seems to really be about (i.e., shareholders and their investments). 

The particular type of materiality matrix advocated by GRI takes the form of a graph (see Figure 1). Figure 2 depicts the same idea in the form of a 2-by-2 matrix. Note the labeling of the axes in both cases.  Along the x axis, we see GRI’s first dimension of materiality; along the y axis, we see the second one.  Figure 2 also includes some interpretation on our part as to how impacts in each quadrant can be understood.

Figure 1

Figure 2

Now while it is true that many mainstream corporate sustainability reports do include materiality matrices, they largely depart from the structure and form suggested by GRI and instead take the form shown in Figure 3.  Somehow and somewhere, someone decided to change the impacts criterion shown on the x axis and replace it with the version shown in Figure 3 (or some variation of it).  Instead of considering impacts on the economy, the environment and society a la GRI, materiality, under the new view, focuses instead, on whether, and to what degree, impacts affect the organization and/or its business goals.  

Exactly when this myopic change was made, and by whom, is unclear to us.  But what is clear is that it amounts to a perversion of the idea of materiality in sustainability reporting, because it essentially cuts out consideration of what are arguably the most material issues: the broad social, economic and environmental impacts of an organization, regardless of how they relate to a particular business plan or strategy.

Let’s turn now to what we believe is a better model.  First, we define materiality as follows:

An impact is material to an organization’s sustainability performance if it affects, or should affect, the well-being of a stakeholder.  A stakeholder, in turn, is anyone whose interests and well-being are affected by an organization’s impacts on vital resources in the world, or whose interests and well-being ought to be so affected given the relationship that exists between them. 

What makes an impact material or not, then, is the effect it has or should have on resources required for stakeholder well-being.  So if we remove consideration of our broad social, economic and environmental impacts and replace it with something else, we deprive ourselves of what is arguably the most important criterion for determining materiality: our broader impacts in the world and their effects on human/stakeholder well-being.

A much better model for what a materiality matrix should look like is shown in Figure 4.  Derived from a school of sustainability theory and practice known as context-based sustainability, the matrix shown is based on the view that materiality is principally about impacts on vital resources in the world, which such resources are required for human well-being.  Indeed, it is the existence of such impacts that determines, at least in part, whom an organization’s stakeholders are, and what the focus of its sustainability program should be.

Figure 4

Notably, then, the x and y axes in our context-based materiality matrix both pertain to impacts on vital resources in the world, but in two different ways.  One, the x axis, differentiates between two different types of beneficiaries of such impacts (those to whom moral or legal duties are owed by the organization to have and/or manage their impacts on vital resources versus those to whom no such duties are owed); and the second axis, the y axis, which deals with impacts that are either discretionary (or merely incidental) or obligatory.  The obligatory impacts stem from non-voluntary duties and obligations owed to stakeholders; the discretionary or incidental ones do not.

What the context-based matrix tells us is that all impacts that fall into the upper-right quadrant must be addressed in a sustainability program and/or report because they pertain to moral or legal obligations an organization has to have and/or manage its impacts on resources of vital importance to stakeholder well-being.  The upper-left quadrant, by contrast, is not applicable, since the very existence of an obligation to have impacts of one kind or another implies that the parties involved are stakeholders to whom duties are irrevocably owed.

The lower-left- and right-hand quadrants pertain to impacts that are fundamentally non-material to an organization’s sustainability performance because they involve impacts that are completely voluntary, if not merely incidental.  What's more, if not having the impacts involved (e.g., deciding not to make a grant) cannot be construed as unsustainable, then the actions or inactions involved cannot be germane to sustainability performance at all, and are therefore immaterial.  Thus, it is only actions and impacts that fall into the upper-right quadrant that must always be included in the scope of a sustainability program (or report), regardless of whether impacts in the other areas are included as well.


Some Closing Thoughts

All: One consolidating thought I'd like to offer before this discussion on materiality matrices comes to a natural close is as follows (this is a guiding principle I go by in my own consideration of competing points of view on what sustainability means, materiality frameworks, etc.):

If it is possible under the dictates of a particular sustainability doctrine for an organization to fully satisfy the performance and/or materiality criteria involved, and yet still be functioning in ways that undermine the existence, integrity or sufficiency of vital resources in the world at levels required to ensure human well-being, then the doctrine itself fails on its face and should be rightly rejected.

By this standard, all of the materiality models shown in Figures 1 through 3 in the article above should be rejected, because they make it possible to leave critical aspects of an organization's sustainability performance out of the picture. SAP's relegation of its impacts on water resources and its waste emissions to the non-material region of its matrix, for example, are a case in point. The same goes for GRI and its decision to leave its impacts on water resources out of its own reports. [Don't look now, but freshwater resources are under stress in the world. Thus, everyone's use of them is material to their sustainability performance!]

In GRI's case, the reason given for leaving their water use out of their reports is that their own impacts are insignificant. But that is not the issue, is it? The issue is that their impacts may or may not be sustainable, no mater how small they are? Indeed, it may be the case that everyone's individual use of water resources is insignificant as a percentage of the whole, but taken together may still add up to collective unsustainability, especially if everyone's individual use is unsustainable no matter how small it is.  Of course, if everyone's individual use was sustainable, collective use would be sustainable, too. Which is why all use of water resources should be held to a test of sustainability these days, no matter how small it is as a percentage of the whole.

Still, GRI's materiality matrix and the variant of it seen in reports like SAP's (similar to what I showed as Figure 3) make such omissions possible, and that is their fatal flaw. The fact that they also allow discretionary and incidental impacts to be included in reports is problematic, too, but not as much as the fact that they allow highly material areas of impact to be left out. To be sure, that is their most contemptible characteristic. And that is why they should be rejected!

Regards, Mark

Different Perspectives

It seems we have a fundamental disagreement, one that stems from our differing perspectives on sustainability itself.

We see sustainability as central to an organization’s ability to thrive in an increasingly complex and interconnected world. As such, a matrix showcasing stakeholder concern vs. company impact rightly focuses attention on the specific environmental, social, and economic concerns related to and influenced by company action. Using these axes also avoids the problem of double-counting impacts, which are (if the analysis is properly conducted) already accounted for through the consideration of impacts on stakeholders and effects on the company. Moral and legal “obligations” are addressed by properly structuring how company impacts are defined.

We find little value or utility in pretending that active engagement with the financial world is irrelevant, in part because more and more investors see environmental, social, and governance issues as material. Rather, we see our responsibility to operate in the world that exists—which, like it or not, still focuses on ROI, profit, EPS, and other markers of financial performance—while constantly promoting a vision of sustainability that fully accounts for societal expectations, ethical considerations, and ecosystem constraints.

Finally, because our approach to managing
for sustainability is holistic rather than siloed, we see the labeling of issues as either “financial” or “nonfinancial” as creating false distinctions. Proper consideration of the financial impacts of ESG issues is a crucial component of smart business decision-making and risk assessment. Referring to ESG issues as “non-financial” negates their very real economic and financial ramifications. Even worse, it perpetuates erroneous thinking and dangerous inaction among organizations who stand to benefit by addressing ESG issues in meaningful ways.

Responding Further to Aleksandra


I agree with you on the importance of sustainability to an organization's ability to thrive, if not simply survive. No disagreement there.

As for the value of axes in matrices that speak to stakeholder concerns and impacts on companies, I have no trouble with including that kind of thing in the development of sustainability programs and reports, but not to the exclusion of the more basic issue of what a company's impacts on vital resources are, or should be, in ways that can affect stakeholder well-being. How else are we supposed to define sustainability standards of performance in a meaningful way?

Call me old fashioned, but what I hope and expect to see in a sustainability report is information about whether or not a company's operations are sustainable, just as I expect to see in a financial report information about whether or not its operations were profitable. This can't be done if sustainability standards of performance are missing. If all the report tells me is what stakeholders' concerns are and how the impacts might affect a company, I may or may not get what I'm looking for. In other words, the content may not be material, because norms, standards or thresholds for performance may be missing.

Consider the following example of how the prevailing form of materiality matrices can lead to problems. Using the type of matrix shown in Figure 3 in my article, let's imagine a case where stakeholder concerns as expressed by, say, customers are high in terms of the kinds of things a company's products have impact on (e.g., energy use), and that the relevance of the same issues to the company's business is also high because of the products it makes.

This immediately takes us way off track because the effects of a company's products on a customer's sustainability performance (or purely intellectual interests) may or may not have anything to do with the company's own sustainability, and yet the type of matrix you seem to favor implies that it does. No amount of attention paid to the social or environmental effects of a company's products or services in the world can substitute for an assessment of the direct effects of its own operations. If the production of my product improves your energy efficiency but contaminates large amounts of freshwater resources when I make it, is my performance sustainable? I don't think so, but your matrix could lead to the opposite conclusion and that is its fatal flaw.

Take a look, for example, at SAP's 2010 report. The upper-right quadrant there is at least half filled with product-solution issues, as if the sustainability of a company's own operations is somehow unimportant. Indeed, thanks to that orientation to materiality, 'Green IT' shows up in the least material part of the same matrix, notwithstanding the fact that SAP is itself an IT company with (presumably) a heavy IT footprint of its own.

To be clear, the primary referent of interest in assessing the sustainability performance of an organization is the company itself and its own operations, not its products or the sustainability performance of its customers when they use them. So the very first consideration in determining materiality is to understand what a company's own direct impacts are on vital resources in the world (as a result of the operations under its own control), and whether or not it has a duty or obligation to have and/or manage impacts on them in a way that can affect stakeholder well-being -- or to have impacts of an entirely different kind that it might not yet be having at all.

Stakeholder input and engagement is obviously important here. And so it is not the y axis in Figure 3 that I object to. Rather, it is the elimination of what GRI defines for the x axis (impacts on vital resources) that troubles me most, as if the social, environmental and economic impacts of a company are not important and are somehow trumped by consideration of what the impact on a company's financial performance is. That is one thing GRI gets right and it should not be lost on us.

Next, to be clear, I don't think engagement with the financial world is irrelevant, if by that you mean engagement with shareholders (at least). As I explained in my last message to you, I think non-financial performance is highly relevant to shareholders. Why? Because poor non-financial performance can put shareholder value at risk. Where we seem to disagree is on how best to report non-financial performance in ways that actually serve stakeholders' interests, including shareholders, instead of undermining them. I want to see such reports tell me whether a company is living up to its non-financial obligations. You seem content to leave that aspect of materiality out.

On the other hand if you agree with me, then you have to allow the admission of the 'obligatory versus discretionary' distinction I've made into the determination of materiality, no matter what else you decide to include. Otherwise you undermine the determination of materiality because you prevent anyone from identifying sustainability standards of performance.

Last, there is nothing silo-like about making a distinction between financial and non-financial impacts. Profits are not the same as gallons of water, ecosystem services, or fair trade conditions, no matter how much you or others would like them to be. A company can be profitable and environmentally sustainable, or just as profitable and environmentally UNsustainable. Separate things call for separate measures. That's not silo thinking; it's common sense.



No help at all I think the

No help at all

I think the 'improved' model matrix presented here by Mark McElroy is no improvement at all, and in fact, significantly muddies the waters even further. What does he propose is the difference between a 'gift' and a 'grant'? How does one define 'obligatory' versus 'discretionary' impacts? (How could a company have discretionary impact on water, for instance - isn't this just a way of prioritizing physical environmental impacts and deprioritizing more difficult, risk-based social ones?) And why develop a 4x4 matrix if one of your quadrants is completely moot?

I think the real lesson here is that a materiality matrix is of use only as an internal decision support tool, and not as something to be included in a report in its own right. A report should be completely transparent as to the thinking behind including or excluding any content, such that a matrix - which incidentally almost never contains any real content, just a description of process - is not needed to rationalize a report.

Responding to Judy's Issues


Thanks for your comments and questions. All very constructive. Let me try to answer them.

First, a gift and a grant might be the same thing, but not necessarily. One could say that all grants are gifts, but not all gifts are grants. Are tickets to the ballgame a grant in the sense that a monetary scholarship or an outright cash award is?

More important is where I placed gifts versus grants in the matrix. Even if you prefer to think of them as being the same thing, some are given to stakeholders to whom no obligation to do so exists, and others are given to stakeholders to whom obligations to have impacts on vital resources DO exist, but not for the impacts involved (e.g., holiday gifts given by a consultant to his or her clients). In any event, neither case is material insofar as a sustainability program or report is concerned because they are entirely discretionary, and so withholding them, therefore, would not be UNsustainable. This is why the discretionary versus obligatory distinction is so important to materiality in sustainability programs. It provides us with minimum standards of performance.

As for the precise definitions of discretionary versus obligatory, I have no special meanings in mind here, only the usual ones. Obligatory impacts are those which a person or company is morally or legally bound to have; discretionary ones are not and instead are voluntarily performed. The point is that sustainability is about performance relative to obligatory impacts, not discretionary ones. Not having obligatory impacts is unsustainable because it puts stakeholder well-being at risk. Not performing voluntary acts, by contrast, is NOT unsustainable because there is no duty to have them in the first place. Again, think in terms of moral and legal obligations to have and/or manage impacts on vital resources (social and environmental) of importance to stakeholder well-being.

As for how a company could have discretionary impacts on water, there are many ways this could occur. A company could do so by conserving its use of water in cases where its use is already sustainable (i.e., so others' needs could be better met). Or take steps to help conserve water resources it does not use or rely on at all. Indeed, there are lots of voluntary things a company could do to have impact on water quality and/or quantity in cases where its own use is already sustainable. And none of this, incidentally, has anything to do with de-prioritizing social impacts. Not sure where that came from.

Finally, I may or may not agree that materiality matrices should not be included in reports. It depends. As long as the reasoning behind how a company determines the content of its sustainability program (and report) is revealed, who cares what form the disclosure takes? What's most important to me is that the basis of materially is disclosed and that programs minimally include treatment of impacts organizations are duty-bound to have based on who their stakeholders are and what their obligations are to have and//or manage their impacts on resources of vital importance to their well-being. Absent anything else in terms of how materiality was determined and what it is, such matrices can do just fine.



PS - Sorry about the N/A quadrant. Hope I didn't break some sacred 2x2 matrix rule. The N/A case is a necessary (and no less valid) consequence of the fact that the other three quadrants rely very heavily on the x and y axes as labeled.

Further Clarifications

Aleksandra, of course I agree that materiality matrices DO serve an important purpose, assuming they're usefully constructed. Where we seem to disagree is on whether or not the current generation of such matrices are usefully constructed. I think they miss the mark for the reasons I've explained.

Further, I disagree that contextual information can be included in the current generation of matrices, since they fail to differentiate between precisely the things we list in the revised matrix I proposed: obligatory versus discretionary or incidental impacts, and impacts that affect stakeholders to whom duties to have such impacts are owed as opposed to not owed. These considerations are vital (or material!) to making materiality decisions and their absence from a materiality matrix is fatal to achieving its purpose.

The water example you provide also does not hold up to scrutiny, in my view, because you continue to express impacts in terms of impacts on a company. Sustainability is not about impacts on a company, it is about impacts on vital resources of importance to stakeholder well-being. So the very name of the x axis you and others refer to argues against it. Throw in a z axis with 'impact on company' as its label, if you like, but let's not confuse 'impacts on company' with 'impacts on resources'.

Further, the water example you gave fits precisely into the upper-right-hand quadrant of the revised matrix I propose, so I think the example is a good one even if the matrix you would have us use in connection with it is flawed.

Next, your appeal to what the financial world may or may not find practical is lost on me. When did we agree that materiality matrices must pass muster with the financial world? I certainly do not think that is the case, and that to put things in those terms is to pander to the shareholder primacy doctrine. I think we're beyond that at this stage, aren't we?

That said, to answer your subsequent question, what makes the model we propose important to shareholders is that the security of their investments depends, in part, on how well a company lives up to its NON-financial duties and obligations, not just its financial ones. Thus, the matrix we propose serves to identify what a company's specific non-financial duties and obligations actually are, and to whom they are owed. Failing to do this puts shareholder value at risk, not to mention the well-being of other stakeholders whose legitimate interests are also affected by what a company does or does not do.



Materiality Matrices DO serve an important purpose

We agree that context-based sustainability is an important goal in both strategic planning and disclosure.

There's no reason, however, that contextual information can't be embedded within the type of materiality matrix most commonly in use today (stakeholder concern on y-axis, impact on company on x-axis)... and thus achieve the goals of identifying key sustainability issues.

For example, water scarcity certainly impacts many companies' operations, and proper planning would take into account total withdrawals as compared to ecosystem carrying capacity, local communities' water needs, etc. ALL as part of the calculation of total impact on the company. The "impact on company" axis bundles contextual information into a total assessment of impact.

Presenting the matrix with a "company" component is quickly understandable, and more importantly, easily linked to business strategy/objectives. Companies seeking to align and integrate sustainability and core business strategy can map their path; analysts can follow it; and progress can be made towards assigning financial valuation to both tangible and intangible aspects of a companies' impact.

The matrix you propose will likely be seen as too impractical or too idealistic for most companies trying to speak the language of the financial world and make the point that sustainability is strategic and relevant. Also, the points you're arguing already seem to be addressed by properly conducted materiality analyses.

As such, we'd be interested in hearing more about how you view your proposed matrix in terms of potential uptake by companies (in the practical course of convincing investors and shareholders of the strategic significance of sustainability) and usefulness in terms of integrated strategic planning and disclosure.

For more on Framework's perspectives on materiality analysis, please see our blog series on materiality and integration:

Use of Materiality Matrices

Many companies now use that "influence on stakeholder decisions" (the Y axis) as a proxy for "significance of economic, environmental and social impacts" (what used to be the X axis in your example). They can do that because among their stakeholders are NGOs and other interest groups that are advocating on behalf of economic, environmental and social objectives. Because of this, the dual axes--one that reflects external impacts and one that reflects company impacts--is a pretty good way of talking about materiality, in my opinion.

Stakeholder Decisions

David, while what you say may be true, we can't have a model that some organizations interpret one way and others interpret another way. That's untenable, and it only highlights the weakness of the prevailing approach: they're too ambiguous and imprecise. Further, if what we're really trying to focus on is impacts that affect stakeholder well-being, why not just say that in a materiality matrix? Indeed, it is my contention that materiality is not about 'decisions and assessments' at all, unless it is about decisions and assessments by stakeholders as to whether impacts on vital resources are affecting (or should be affecting) their well-being. Again, why not just say that?

Mark W. McElroy, Ph.D. is the founder and Executive Director of the Center for Sustainable Organizations and the original developer of the Context-Based Sustainability method. He is also co-founder of Thomas & McElroy LLC, creators of the MultiCapital Scorecard, and… [Read more about Mark McElroy]

  Sign up for SB Newsletters
Get the latest personalized news, tools, and virtual media on a wide range of sustainable business topics in your inbox.


User login

Engage the community