How ESG Exposure Research Services
Built a Traceable Data Feed for an EU Ratings Provider
An EU Exposure Metrics Data Vendor Supporting Institutional Screening
The client operates across Europe and delivers ESG risk exposure data to banks, asset managers, and investors focusing on sustainability. The client’s datasets support business involvement screening by identifying corporate activities that may conflict with portfolio policies, exclusion criteria, or sustainability mandates. Their proprietary ESG ratings system records disclosed activities, revenue-based exposure, geographic exposure data, and classification details for every covered company.
Building a Source-Traceable Revenue-Based Exposure Dataset for 10,000+ Companies
The project covered company-level ESG data research and data collection from multiple sources for more than 10,000 companies across 10+ exposure metrics. Our ESG specialists had to identify specific activities (as per client guidelines) and quantify the associated share of company revenue. Each finding needed traceable evidence, not general sector classification or qualitative exposure research. The client also required a team that could internalize its methodology, implement evolving SOPs, and scale controversial activity screening without disrupting delivery.
The engagement required:
- Methodology adoption: Convert the client’s Metric → Activity → Evidence → Revenue hierarchy into a consistent analyst workflow and apply every SOP revision across active research.
- Evidence-based classification: Collect and cross-check ESG disclosure data from corporate filings, official business portfolios, and product or service catalogs before assigning activity classifications.
- Jurisdictional analysis: Capture sector and geographic exposure data when jurisdictions classified gambling, cannabis-adjacent products, or defense exports differently.
- Revenue estimation: Calculate revenue-based exposure percentages from disclosed segment revenue, using documented asset-based or product-based proxies when standalone figures were unavailable.
- SOP governance: Maintain a version history for changing procedures, reassess affected records, and preserve production volume and source-level traceability during every update.
- Data delivery: Publish validated company-level ESG data in the client’s proprietary rating system for use as due diligence data feeds in downstream workflows.
Categories for exposure metrics data collection (as per client guidelines):
- Fossil Fuels and Related Energy
- Nuclear Energy
- Gambling and Leisure Industry
- Oil-Sands Production and Participation
- Animal Testing and Animal Cruelty
- Arctic Oil and Gas
- Adult Entertainment Industry
- Coal Mining Production and Participation
- Contraceptives and Abortifacients
- Alcohol and Other Drugs
- Conventional Weapons Production and Participation
Preserving Data Integrity as Methodology, Coverage, and Team Size Evolved
The engagement required a complete audit trail while the client’s methodology kept changing. Our team had to preserve evidence lineage, classification accuracy, and delivery cadence while the client revised its SOPs.
Reframing ESG Analysis around Revenue
Our analysts had more than five years of experience with related ESG datasets, but exposure research required a different reasoning model. Traditional ESG ratings commonly evaluate company performance against broad environmental or social topics, often using qualitative assessment. The client’s methodology instead required activity-level revenue-based exposure. Analysts had to quantify the revenue associated with each defined activity using disclosed evidence and reproducible calculations. Broad sector labels or qualitative judgment could not support the final classification.
Reapplying Revised Rules to Completed Research
The client revised its SOPs repeatedly while resolving complex structures and incomplete revenue disclosures. New guidance covered REITs with mixed property portfolios, conglomerates with minor exposed segments, and companies requiring asset-based revenue proxies. Our team had to identify completed records affected by each SOP revision and back-test them using the updated calculation rules. Analysts also documented resulting differences and confirmed interpretations with the client’s research leads. We had to implement these changes without reducing weekly throughput or weakening the audit trail.
Expanding Capacity without Methodology Drift
The engagement began as a three-month pilot program managed by five analysts. After validating the research model, the client rapidly expanded the company universe and the number of exposure categories. New analysts could not rely solely on conventional ESG research practices because the project required precise activity-level classification. Scaling therefore depended on structured onboarding, multilevel peer review, and strict SOP version tracking. This discipline enabled activity-based risk profiling to expand without introducing quality or classification drift.
Constructing Revenue-Based Exposure from Fragmented Evidence
Revenue attribution became difficult when controversial activities were embedded within broader business lines. Financial reports rarely separated airline alcohol sales, casino income within hospitality REITs, or contraceptive revenue within diversified pharmaceutical companies. Analysts therefore combined company websites, official menus, business portfolios, segment disclosures, and asset-based proxies. For a gaming REIT, the proportion of gaming-related properties could support the exposure calculation. Every estimate still had to remain conservative, reproducible, and traceable to its underlying sources.
Building Audit-Ready Exposure Datasets through Methodology Governance and Multi-Level Review
Before scaling our ESG exposure data collection services, we ensured that every analyst understood the client’s methodology. The team then applied approved category defaults, calculation rules, and the latest SOPs. This kept each company-level ESG record linked to its evidence, revenue formula, and methodology version.
Research Model Validation
The engagement began in June 2025 with a three-month proof of concept managed by five full-time ESG analysts. We selected each analyst from our wider research pool for proven experience with disclosure-based ESG data research. The initial assignment covered selected companies across a restricted set of exposure metrics.
The pilot tested the complete workflow from metric identification through revenue calculation and quality review. The team resolved methodology questions before introducing larger production volumes. The team also developed category reference materials, working templates, and formal research-to-QA handoff procedures.
Record of SOP Changes
We established a formal change register for the client’s evolving methodology. We documented every SOP update, had peers review it, and discussed it with the client’s research leads before implementation. This process ensured that all analysts received the same interpretation.
When an update changed the calculation logic for an activity class, we mapped it to previously completed records. We recalculated only affected records under the new rule, reconciled the resulting differences, and documented them. Records outside the update remained intact. This approach preserved historical accuracy and made the governing SOP version identifiable for every record.
Evidence Mapped to Activity Codes
The company-level research protocol followed the client’s Metric → Activity → Evidence → Revenue hierarchy. It combined AI-assisted discovery with mandatory primary-source verification:
- Our analysts used ChatGPT, Perplexity, Claude, and Gemini to locate potential evidence within ESG disclosures, corporate filings, and published reports.
- Analysts checked every AI-generated or manually sourced lead against an existing, accessible document, filing, webpage, or cited page.
- Confirmed leads entered the active research queue. Excluded unverified leads and recorded them to measure AI research reliability over time.
- Disclosed activities were assigned to mutually exclusive codes, including distinctions between exploration, refining, distribution, Production, and Participation.
- We examined subsidiaries, joint ventures, and equity interests to capture business involvement not visible at the parent-company level.
- Each classification included at least one primary citation, with the source document, page, and publication date preserved in the audit trail.
Revenue Exposure Quantification
Corporate filings did not always report activity-level revenue separately. Although disclosed segment revenue provided the preferred calculation basis, niche product lines were often grouped within broader business segments. Our analysts moved beyond qualitative exposure research by combining financial disclosures with verified company and product evidence. Documented proxies replaced unsupported assumptions, keeping every quantitative estimate traceable to its sources.
For example, a diversified pharmaceutical company might report total healthcare-segment revenue without isolating contraceptive product sales. Analysts verified the relevant products through official portfolios and applied the client’s documented product-based proxy to the associated segment data. For REITs and holding companies, they calculated the proportion of exposed assets and applied that ratio to the corresponding revenue stream. Every estimate was conservatively calculated, explicitly labeled, and supported by a formula that reviewers could retrace end to end.
Two Independent Review Gates
Every completed record passed through a two-level data quality review. Peer reviewers checked activity codes, classification accuracy, and missing evidence. Senior reviewers examined methodology compliance, revenue formula integrity, edge-case treatment, and alignment with the latest SOP.
This review structure maintained a 95%+ QA pass rate while coverage expanded to thousands of companies. The same quality level continued as the research team grew from five to thirteen analysts.
Capacity Expansion in Controlled Cohorts
We added analysts in small cohorts as the client increased its coverage targets. Senior team members led structured onboarding so each cohort understood the methodology before assuming production responsibilities. New analysts also learned category defaults, current SOPs, evidence requirements, and revenue calculation rules.
Across more than ten months of engagement, the team expanded from five to thirteen analysts. Despite this capacity increase, average throughput remained above 100 companies per analyst each month.
Project Outcomes
Institutional Screening Coverage across 10,000+ Companies
We provided an accurate, activity-based ESG data feed supported by comprehensive exclusion lists and portfolio-level screening for the client’s institutional customers.
Dataset Quality Sustained Above 95%
Quality remained stable as company and category coverage increased. This let the client’s researchers focus on methodology-sensitive cases rather than routine remediation.
More than 10 Exposure Categories Operationalized
Production coverage included fossil fuels, alcohol, gambling, animal testing, and other controversial activities aligned with the client’s institutional screening criteria.
Team Expanded over 10 Months
The analyst team grew from 5 to 13 in 10+ months. This expansion supported rising coverage targets without measurable declines in QA rates or individual productivity.
Build a More Defensible ESG Exposure Dataset
Planning to build or expand an ESG exposure dataset? We provide ESG exposure research services covering company-level ESG data collection, controversial activity screening, revenue-based exposure analysis, and version-controlled sustainability data management. Discuss your requirements with us and get a free sample.