Hit enter to search or ESC to close
As ESG performance becomes ever more central to operational success, the key role sentiment analysis plays in tracking it needs to be recognised.
The current business landscape, sculpted by the ongoing pandemic, as yet unchecked climate change, and the rise in popular movements demanding corporates honour social contracts, requires that every organisation has an ESG strategy in place. The rise in visibility of environmental, social and governance (ESG) issues have made the ability to measure ESG performance as vital as tracking profit and loss – not least due to the acknowledgement that the two are indelibly linked.
ESG measurement, however, is more complex than balancing the books.
The challenges range from the breadth of the subjects, speed with which new topics emerge and evolve, the fickle nature of stakeholder perceptions in relation to ESG issues, a lack of consistent regulation, and consequent proliferation of voluntary reporting and self-disclosure around ESG performance.
Other impediments to effective, consistent ESG measurement include the trap of focusing only on issues that are easier to enumerate or monetise, which may result in missing key factors that matter deeply to stakeholders, and artificially inflating ESG performance.
Tracking inputs instead of outcomes is another stumbling block. Just because an organisation is doing well on stated diversity targets or sustainability goals, doesn’t automatically mean it has created a more inclusive culture or reduced its environmental impact.
Subjectivity also plays a huge part. What one stakeholder considers a vital element of ESG performance may be irrelevant to another. And their perception of what ranks as positive performance will also differ. For one, offsetting carbon emissions without impacting profits will be enough. For another, nothing short of converting all business operations to emission-free renewable energy will classify as success.
In order to aspire to any level of accuracy, ESG measurement requires a formula that doesn’t obscure insight, misdirect from key topics, or misrepresent performance. A central component in that formula is sentiment analysis.
Sentiment analysis uses artificial intelligence and natural language processing (NLP) to establish the disposition of content towards its subject matter. For a global organisation wanting to understand the varying attitudes of its stakeholders to its business operations, this represents the gathering, processing and interpreting of millions of pieces of publicly available data, garnered from sources as diverse as newspaper reports, television and radio broadcasts, social media posts, press releases, AGM minutes, online reviews, employee surveys and government debates.
Sentiment analysis works by recognising and quantifying positive or negative feeling at a document, sentence and phrase level. This granular approach can also filter by author, topic, sector and company. To do this, the programme has to comprehend the meaning of language, identify individual voices, link subjects and interpret irregular or vernacular terms.
Machine learning enables sentiment analysis solutions to constantly refine their ability to pick up nuance in the opinions it encounters. The continuous development of semantic dictionaries and internal libraries of related phrases gives ever more accurate sentiment identification. This technology raises sentiment analysis several degrees above a rules-based system interpreting standard language, to one that can map and interpret the illogical nature of human communication.
Carried out at a stakeholder level, sentiment analysis can surface how different groups prioritise and respond to specific topics. The reaction to announced job cuts, for example, will differ starkly between employees, different levels of the business hierarchy, customers, and shareholders.
This is why the synthesis between sentiment analysis and ESG measurement is fundamental to the accuracy of the latter.
Sentiment analysis will become an increasingly significant element of ESG intelligence, because it allows organisations to understand how stakeholders feel about their achievements in the ESG sphere. This represents a mirror in which is reflected actual ESG performance.
And it can do it in real time. This is crucial in a rapidly evolving arena in which new topics are constantly emerging and opinions shifting. With stakeholder priorities and expectations in a constant state of flux, speed is essential for both risk management and in order to exploit associated opportunities.
On its own, primary research into how stakeholders feel on any given topic is both too slow and limited in the scope with which it can be performed. To comprehend myriad, rapidly changing views, businesses need to infer sentiment from the data that already exists.
The structure of the ESG landscape also means that the map looks significantly different between sectors. Sentiment analysis can provide a company-specific and broader industry view of how an organisation is performing, what others in the sector are doing, and how performance is changing over time.
At the forefront of sentiment analysis adoption in ESG measurement is the sustainable investing fraternity, interrogating ESG sentiment data to guide investment decisions. With the global market for ESG investments forecast to reach $45tn by 2025, ESG measurement is a pressing priority for investors. To create a viable investment strategy, they need a solution that can keep pace with the speed at which stocks and shares are traded. Sentiment analysis is the answer for both institutional investors – and for any company looking to operate at a similar speed and with similarly accurate levels of quantification.
While it has yet to become the dominant solution in the field, the fact is, it is simply not feasible to accurately measure ESG performance without employing machine learning and NLP-based sentiment analysis. There is no alternative research method that can accurately provide insight into stakeholder perception at a velocity to match the evolution of the issues.
The likely future developments in sentiment analysis make this inevitable. As technology develops, the next generation of ESG sentiment calculation will encompass:
Be part of the Stakeholder Intelligence community