The Company Peer Group using Similarity Analysis Blueprint leverages S&P Global data on publicly listed companies to derive similarity scores. First, a similarity score is calculated using quantitative financial and market data. Next, a second score is calculated using business descriptions. The two scores are combined to generate a final score, which can be used to rank companies and construct a Company Peer Group.

This workflow utilizes natural language processing (NLP) to convert sentences into high-dimensional vectors (embeddings) to capture the semantic meaning of these sentences. The use of a pre-trained, open source model reduces compute costs of the workflow.

The workflow can be utilized to build a company’s peer group for general market analysis, portfolio analysis, or peer research and due diligence. It is also applicable in competitor intelligence analysis, as demonstrated in the Competitor Intelligence Blueprint. Additionally, it provides foundational elements for a more advanced peer group creation workflow.

This Blueprint includes:

• S&P Capital IQ Financials: Standardized and As Reported financials data for global public companies as well as thousands of private companies and private companies with debt.
• S&P Capital IQ Estimates: Comprehensive global estimates based on projections, models, analysis and research. Available as global, international or North American company packages.
• Market Data: Currency and split-adjusted pricing data across open, close, high, low, volume, adjustment factors, shares outstanding, volume-weighted average price and other data points.
• Company Intelligence: Profiles of public and private firms worldwide including company contact information, competitors, financial auditors and summary or detailed-level business descriptions.
• GICS®: GICS is the standard recognized by market participants worldwide for global sector and industry classifications.
• S&P Capital IQ Workbench: A web-based notebook environment that allows you to perform data analysis or build models using collaborative notebooks that support multiple languages, built-in data visualizations, and automatic versioning.