Our supplier risk assessment methodology covers the following aspects:
Covers financial performance and stability of a supplier on the basis of profitability, solvency, liquidity and sponsor/key investor type
Covers operational challenges arising from factors such as size and scale, business concentration, competition, management quality, security, product quality, geographic dependencies, supply and demand, and impact of technology
The geographic risk indicators are explained below:
Includes ethical and governance challenges and overall business conduct
Covers reputational and other risks related to labour health and safety, labour compensation, labour rights, sustainability of operations (or production) and product safety
Covers internal risk signals based on the proposed indicators below (to be finalised in consultation with the client).
Overall, our risk assessment covers 49 external and 25 internal risk parameters. The top 4 risk types (Financial, Operational, Legal & ethical, and Human & environmental) cover topics that are relevant in today’s world and meet the needs of leading companies globally.
The Legal & ethical and Human & environmental risk types in particular cover all relevant topics for ESG risk assessment and management; these 2 risk types and the underlying parameters are aligned with globally recognised standards, such as the United Nations Global Compact, World Federation of Exchanges, Global Reporting Initiative and Sustainability Accounting Standards Board.
The illustration below captures our risk assessment and scoring process.
The first step is to collate risk event feeds from different sources. We cover +100,000 news media sources, top social media platforms, regulatory websites, company registrars and exchange commissions, and all major national and global statistics organisations.
Next, we use advanced analytics and machine learning algorithms to filter, classify and rate risk event feeds. Sentence-BERT and LightGBM algorithms are used to filter junk/non-relevant information feeds, while deep learning models are used to classify and score events. Our machine learning models are regularly trained to sharpen risk classification and identification.
To support the risk scoring process, we also analyse supplier financial information and key country indicators, which are then fed into the system. Our clients have the option to include licensed third-party data and internal supplier SLAs/KPIs to get a more holistic view of their suppliers’ risk.
Our robust risk scoring model generates risk scores that provide a holistic view of risk associated with suppliers. The model comprises the following elements:
The overall risk score is a dynamically weighted average of adjusted risk scores for all sub-risks and risk types. Our risk scoring formula is ‘dynamic’ in the sense that it automatically adjusts weights to show ‘real’ risk when a singular risk event (e.g., reported criminal investigations or labour rights violations) has a material implication for the supplier’s business. This adjustment is required to ensure that the impact of disruptive events is properly captured and projected, and not subdued due to use of weights across multiple parameters.
Our risk specialists are experienced analysts who have been trained to validate risk signals, assess their impact and provide a brief summary of risks associated with a supplier. The succinct risk summary helps clients quickly make sense of different risk indicators and event feeds, thus ensuring meaningful insights.