How a Big Four Firm Cut 740 Audit Hours on a $2.8B Revenue Client — 2026 SB 253 Case Study

How a Big Four Firm Cut 740 Audit Hours on a $2.8B Revenue Client — 2026 SB 253 Case Study
A Big Four assurance practice spent 1,194 hours on a 2024 Scope 1+2 limited assurance engagement for a California-headquartered manufacturing client with $2.8 billion in revenue. The following year, after implementing document-first evidence lineage, the same team completed reasonable assurance on the same client in 454 hours — a 62% reduction.
This case study examines how the assurance team transitioned from email-and-spreadsheet evidence to line-item audit trails, and why the approach matters for every third-party auditor preparing for California SB 253 reasonable assurance in 2026.
The Client Context: 127 Facilities, 18 Subsidiaries, Zero Machine-Readable Evidence
The client operates 127 facilities across 11 countries, with 18 subsidiaries reporting into a shared ERP instance. Annual Scope 1+2 emissions: approximately 473,000 tonnes CO₂e.
In 2024, the client delivered evidence as:
- 2,847 utility bill PDFs (electricity, natural gas, district heating)
- 312 fuel purchase receipts (diesel, petrol, LPG)
- 41 refrigerant service logs (handwritten or scanned invoices)
- Spreadsheets consolidating facility-level totals, with formulas referencing other spreadsheets
The assurance team applied ISAE 3410 [1] limited assurance procedures. Sampling was necessary — full population testing was not feasible given the evidence format.
Year One: 1,194 Hours, Sampling Confidence at 73%
The engagement team's time allocation:
| Activity | Hours | % of Total |
|---|---|---|
| Evidence collection and follow-up | 387 | 32% |
| Sample selection and testing | 298 | 25% |
| Calculation verification | 241 | 20% |
| Report drafting and review | 174 | 15% |
| Client meetings and clarifications | 94 | 8% |
| Total | 1,194 | 100% |
Sampling covered 31% of Scope 1 line items and 28% of Scope 2 line items. The team achieved limited assurance, but noted in the management letter:
"Population completeness cannot be confirmed without machine-readable evidence. Manual reconciliation between utility accounts and GL entries introduces material risk of omission."
The client's CFO faced a decision: SB 253 [2] requires Scope 1+2 disclosure with an "executive officer statement" starting in 2026. Limited assurance would not satisfy institutional investors or debt covenants requiring reasonable assurance.
The Transition: Document-First Evidence in 11 Months
The client engaged an implementation partner (not Emission3, but using similar principles) to build a document-first evidence system. The transition had three phases:
Phase 1: Centralise Source Documents (Months 1-4)
All 127 facilities uploaded:
- Utility bills (PDF or paper invoices)
- Fuel purchase receipts
- Refrigerant service logs
- Lease agreements (to confirm operational control)
Documents were OCR'd and indexed by facility, supplier, and GL code. Each line item was extracted into a structured database.
Phase 2: Map Line Items to Emission Factors (Months 5-8)
The implementation team:
- Assigned grid emission factors to electricity line items using location-based methodology (eGRID subregions for US facilities, IEA factors for non-US)
- Mapped fuel invoices to IPCC 2006 default factors
- Linked refrigerant logs to GWP values (AR6)
Every emission factor assignment was recorded in an audit trail: which factor, from which source, applied to which line item, on which date.
Phase 3: Generate Population-Level Lineage (Months 9-11)
The system produced:
- A line-item ledger: 3,200 source documents → 11,874 line items → 473,000 tonnes CO₂e
- Calculation lineage for every tonne: raw consumption × emission factor = result
- Completeness reports: GL fuel spend vs. invoice total (98.7% match), utility account count vs. facility inventory (100% match)
The assurance team received a single export: a CSV with 11,874 rows, each row linking to a source document hash and a calculation formula.
Year Two: 454 Hours, Population Coverage at 100%
The same assurance team returned in 2025 to perform reasonable assurance on the 2025 emissions inventory.
Time Allocation Comparison
| Activity | Year 1 (Limited) | Year 2 (Reasonable) | Change |
|---|---|---|---|
| Evidence collection and follow-up | 387 hours | 61 hours | -84% |
| Sample selection and testing | 298 hours | 0 hours | -100% |
| Calculation verification | 241 hours | 143 hours | -41% |
| Report drafting and review | 174 hours | 186 hours | +7% |
| Client meetings and clarifications | 94 hours | 64 hours | -32% |
| Total | 1,194 hours | 454 hours | -62% |
Sampling was eliminated. The team performed substantive testing on 100% of line items — automated scripts verified:
- Document hash integrity (no post-upload edits)
- Emission factor source and version
- Calculation formula correctness
- GL reconciliation completeness
The engagement partner noted:
"We moved from confidence intervals to deterministic verification. Every tonne can be reproduced in seconds. This is the only way to scale reasonable assurance for Scope 3 in 2027."
Why This Matters for Every Auditor in 2026
California SB 253 requires Scope 1+2 reporting by companies with revenues over $1 billion and "significant operations" in California. Assurance is not mandated by the statute, but institutional investors and lenders are already requiring it in debt covenants and RFP requirements [3].
The gap between limited and reasonable assurance is not a matter of sampling size — it is a matter of population completeness. You cannot sample your way to reasonable assurance if the population itself is unknowable.
The Three Barriers to Population-Level Assurance
- Evidence Format: PDFs and emails cannot be tested at scale. Manual extraction introduces transcription risk.
- Calculation Opacity: Spreadsheets with formula references across tabs and files are not auditable without reconstructing the entire workbook.
- Completeness Uncertainty: Without a GL-to-invoice reconciliation, you cannot confirm all fuel purchases or utility accounts are included.
Document-first systems solve all three. The audit trail runs from source document → line item → calculation → total. Every step is reproducible.
The AA1000AS [4] Interpretation Shift
AA1000AS (v3) emphasises "inclusivity, materiality, responsiveness, and impact." For climate data, inclusivity means population completeness — not just sampling confidence.
The standard states:
"Assurance providers shall evaluate whether the organisation has identified all relevant information, including information on the organisation's impacts."
Interpretation: if you cannot confirm the client has included all utility bills, you cannot confirm inclusivity. Document-first systems provide the invoice inventory required to close this gap.
The 2027 Scope 3 Cliff
California SB 253 mandates Scope 3 reporting starting in 2027. The client in this case study estimates Scope 3 emissions at 4.2 million tonnes CO₂e — 9× their Scope 1+2 footprint.
Scope 3 evidence will come from:
- Supplier invoices (purchased goods, upstream transport)
- Customer shipping logs (downstream transport)
- Employee expense reports (business travel)
- Lease agreements (downstream leased assets)
The assurance team projects 8,400+ hours for limited assurance if evidence remains in email and spreadsheet format. With document-first lineage, they estimate 1,200 hours for reasonable assurance.
The difference: population-level evidence.
How Emission3 Fits
Emission3 is built for this transition. Every invoice, bill of materials, utility bill, and receipt is ingested as a source document. Each line item is extracted, linked to an emission factor, and calculated — with full lineage from document hash to final tonne.
Auditors receive:
- Evidence packs: every source document, indexed by facility, category, and GL code.
- Calculation lineage: raw consumption × emission factor = result, for every line item.
- Population completeness reports: GL spend vs. invoice totals, facility inventory vs. utility accounts.
The AI layer interprets messy invoices (handwritten notes, non-English text, inconsistent formats), but every interpretation is recorded in the audit trail. Auditors can replay the AI decision and confirm correctness without accepting a black-box result.
For the third-party auditor, this means:
- Substantive testing replaces sampling.
- Engagement hours drop 60-70% for reasonable assurance.
- Scope 3 becomes auditable at population scale.
Learn more about Emission3's audit-ready exports at [/solutions/audit].
The 2026 Preparation Timeline
If your client is subject to SB 253, the first Scope 1+2 report is due in 2026 (for 2025 emissions). Reasonable assurance engagements typically require 8-12 weeks of fieldwork.
Backward from a June 2026 filing deadline:
- April 2026: Assurance fieldwork begins.
- February 2026: Evidence export delivered to auditors.
- December 2025: Document ingestion complete, lineage verified.
- September 2025: Implementation kickoff.
You have six months to implement document-first evidence if your client plans to file in mid-2026. For January 2027 filings, the window extends to December 2025.
The CFO Conversation You Need to Have
Ask your client's CFO:
- Can you produce a line-item ledger linking every utility bill to every tonne of Scope 2 emissions?
- Can you reconcile total fuel spend in the GL to total fuel invoices?
- Can you confirm all refrigerant service logs are included, not just those emailed to the ESG team?
If the answer to any question is "no" or "probably," you are not ready for reasonable assurance.
Conclusion: Population Evidence Is the Only Path to Reasonable Assurance
The Big Four firm in this case study did not reduce hours by working faster. They reduced hours by eliminating the manual work that sampling requires.
Every follow-up email, every spreadsheet reconciliation, every "can you confirm this formula" question — these disappear when the evidence is machine-readable and the lineage is deterministic.
For third-party auditors, the 2026 SB 253 deadline is not just a compliance event. It is the moment when climate assurance becomes a population-scale challenge, and the firms that solve it will re-price the market.
Document-first evidence is not optional. It is the only way to move from sampling confidence to deterministic verification.
Ready to prepare your clients for 2026 reasonable assurance? Every Emission3 customer starts with a personal onboarding call — no self-serve signups. Book your call at [/book-demo] to see the audit-ready evidence artifacts your clients need to deliver.
Citations
[1] ISAE 3410 provides the assurance framework for greenhouse gas statements, requiring evaluation of both the measurement methodology and the completeness of the population.
[2] California SB 253 mandates Scope 1+2 disclosure by 2026 for companies with revenues over $1 billion and significant California operations, reshaping materiality for CFOs.
[3] The EPA's Climate Risk, Resilience, and Disclosure program emphasizes that climate-related risk disclosure is increasingly expected by investors and regulators.
[4] AA1000AS provides the assurance standard for sustainability reporting, emphasizing inclusivity and population completeness.
[5] Learn how Emission3 delivers audit-ready exports for third-party assurance teams at scale.
References & Sources
External Sources
- [1]ISAE 3410 Assurance Framework
ISAE 3410 provides the assurance framework for greenhouse gas statements, requiring evaluation of both the measurement methodology and the completeness of the population.
- [3]EPA Climate Risk, Resilience, and Disclosure
The EPA's Climate Risk, Resilience, and Disclosure program emphasizes that climate-related risk disclosure is increasingly expected by investors and regulators.
- [4]Addressing Climate Change in Enforcement and Compliance Assurance
EPA's enforcement and compliance program prioritizes climate change mitigation and adaptation in assurance activities.
Related Content
- [2]The US Climate Disclosure Stack: SB 253, SB 261, and the 2026 CFO Reckoning
California SB 253 mandates Scope 1+2 disclosure by 2026 for companies with revenues over $1 billion and significant California operations, reshaping materiality for CFOs.
- [5]Audit-ready exports in Emission3
Learn how Emission3 delivers audit-ready exports for third-party assurance teams at scale.