Board reporting scramble: SAP & Salesforce -> Snowflake/dbt to approved Power BI snapshot

Overview

Board reporting at a global manufacturer had become a scramble of last-minute spreadsheet merges across SAP, Salesforce, and regional email attachments. Definitions drifted, cut-off times were unclear, and leadership debated which spreadsheet was correct rather than the substance of decisions. Intelligex built an automated reporting pipeline using Fivetran, dbt, and Snowflake, feeding a governed Power BI template. A formal approval gate with Finance and Strategy locked each reporting snapshot before decks were exported and published to Nasdaq Boardvantage. Executives worked from consistent evidence with fewer handoffs, and discussions centered on decisions because every slide drew from the same controlled dataset.

Client Profile

• Industry: Global discrete and process manufacturing
• Company size (range): Multinational enterprise with regional business units
• Stage: Established public company
• Department owner: Strategy, Analytics & Executive Leadership (Office of the CFO and Corporate Strategy)
• Other stakeholders: Finance and FP&A, Regional General Management, IT/Enterprise Data, Sales Operations, Internal Audit, Investor Relations, Legal & Compliance

The Challenge

Each board cycle triggered a familiar rush. Finance exported actuals and forecasts from SAP, Sales Operations pulled pipeline and bookings from Salesforce, and regional teams emailed spreadsheets with local adjustments and commentary. Analysts reconciled conflicting versions of revenue, backlog, and working capital by hand. Currency conversions, calendar cut-offs, and business unit roll-ups changed during the scramble, so the “final” deck reflected a stitched story rather than a coherent snapshot.

Tooling was already in place: SAP for enterprise resource planning, Salesforce for commercial data, a cloud warehouse used by IT for other workloads, and Power BI for internal reporting. The gap was not a lack of technology; it was the absence of a governed pipeline, shared definitions, and a publish process tied to formal approval. Decks for Nasdaq Boardvantage were assembled manually from files on network drives and email chains, with no single place to confirm which metric definition or exchange rate applied to each slide.

Why It Was Happening

Data and process were fragmented. SAP and Salesforce tracked different slices of the truth with no shared calendar for reporting cut-offs. Regional teams operated on local schedules and currencies, then emailed spreadsheets that were not tied to a central data contract. Analysts wrote one-off formulas to make numbers “match,” which introduced new discrepancies when inputs changed. A late email could override a prior decision without a traceable rationale.

Governance was bolted on at the end. There was no pre-publication snapshot that Finance and Strategy could sign off on, no versioned metric catalog, and no controlled workflow to ensure slides drew from a single source. Publishing to Boardvantage happened after the firefight, which forced leaders to trust that the deck matched the latest reconciliations rather than being able to verify it in a system of record.

The Solution

We built an automated, governed reporting pipeline that aligned data sourcing, metric definitions, and deck production. Fivetran moved data from SAP and Salesforce into Snowflake on a schedule aligned to the board calendar. dbt standardized definitions for revenue, bookings, backlog, and operating metrics, applied currency and calendar logic, and produced conformed tables for consumption. A locked Power BI template rendered the board package from a timestamped dataset. A dual-approval gate in the workflow required Finance and Strategy sign-off on the snapshot before the deck was exported to PDF and published to Nasdaq Boardvantage. No core systems were replaced; the orchestration unified them into a predictable process.

• Automated ingestion from SAP and Salesforce to Snowflake using Fivetran connectors and file feeds for regional submissions
• Standardized modeling and tests in dbt with a documented metric catalog, currency conversion logic, and cut-off alignment
• Schema and data quality checks (completeness, referential integrity, duplicate prevention) with alerting to Finance and Strategy when anomalies surfaced
• Snapshot orchestration that freezes a reporting version after checks pass, creating an immutable dataset for that cycle
• Governed Power BI template with parameterized snapshot selection, locked visuals, and consistent narrative layout (Power BI)
• Approval workflow capturing CFO and Head of Strategy sign-off before deck export and publication
• Automated export to PDF and delivery to Nasdaq Boardvantage with version metadata and access controls
• Role-based permissions and row-level security for pre-read views, aligned to Internal Audit guidance
• Operational dashboards for data freshness, test results, and publication status to reduce status chasing

Implementation

• Discovery: Mapped the board deck structure, current slide sources, and the cut-offs used by Finance, Sales, and regions. Collected definitions for core metrics and their known exceptions. Cataloged SAP and Salesforce fields, regional files, FX rate sources, and the handoffs leading to Boardvantage.
• Design: Defined the metric catalog and data contracts for tables feeding the deck. Established the reporting calendar and snapshot policy. Designed dbt models, tests, and lineage. Planned the Power BI template structure and the approval checkpoints. Documented the publication runbook and rollback plan.
• Build: Configured Fivetran connectors and file ingestion for regional submissions. Built dbt staging and conformed models with validations, currency conversion, and calendar alignment. Implemented anomaly alerts. Created the governed Power BI template, parameterized snapshot selection, and wired an approval workflow that gates export and publication.
• Testing and QA: Ran parallel cycles comparing pipeline outputs to legacy spreadsheets. Reconciled differences back to sources, refined metric rules, and tuned cut-off logic. Tested failure scenarios and deck regeneration from the same snapshot to verify immutability.
• Rollout: Launched in observe-only mode, producing a complete deck from the pipeline alongside the legacy process. After stakeholders validated definitions and layout, activated the approval gate and made the automated deck the system of record. Kept a manual contingency path while teams built confidence.
• Training and hand-off: Delivered playbooks for Finance and Strategy on snapshot approvals, for analysts on metric stewardship, and for IT on pipeline operations. Documented the metric catalog and change control. Established a cadence for pre-board reviews and a human-in-the-loop process for late exceptions.

Results

Board materials drew from a single, controlled snapshot. Finance and Strategy agreed on definitions before the deck left the building, and the same dataset powered every slide and appendix. Analysts stopped stitching spreadsheets and instead focused on analysis and narrative. When executives asked for a drill-down, the figures reconciled to the governed template and the conformed tables behind it.

Handoffs reduced and decisions accelerated because the conversation started from shared facts. Regional contributions flowed into the pipeline through defined channels, with currency and calendar logic applied consistently. Audit readiness improved; each published deck carried a traceable record of its data version, checks performed, approvers, and the specific template used. If a clarification was needed, teams regenerated the deck from the same snapshot rather than re-editing slides.

What Changed for the Team

• Before: Spreadsheet merges and email attachments defined the “final” numbers. After: A governed snapshot in Snowflake fed a locked Power BI template.
• Before: Definitions varied by function or region. After: A metric catalog in dbt aligned revenue, bookings, backlog, and operating metrics.
• Before: Cut-offs and FX logic changed midstream. After: Calendar and currency rules were encoded and consistently applied.
• Before: Publication required manual assembly and checks. After: An approval gate captured Finance and Strategy sign-off before automated export to Boardvantage.
• Before: Questions triggered manual reconciling. After: Decks and drill-downs referenced the same conformed tables with an audit trail.

Key Takeaways

• Automate the data path to the board deck; manual merges create drift and erode trust.
• Define a shared metric catalog and encode it in the transformation layer so every slide reflects the same rules.
• Freeze each reporting cycle with a snapshot and require formal sign-off before publication.
• Use a governed BI template to remove layout debates and keep the narrative focused on decisions.
• Keep core systems—SAP, Salesforce, warehouse, BI, and the board portal—while layering orchestration, approvals, and auditability.

FAQ

What tools did this integrate with?
Data flowed from SAP and Salesforce into Snowflake using Fivetran. Transformations, tests, and the metric catalog ran in dbt. A governed Power BI template rendered the deck from a timestamped dataset, and the approved deck was exported and published to Nasdaq Boardvantage. Alerts and approvals were coordinated through the client’s existing collaboration tools.

How did you handle quality control and governance?
Quality was enforced through dbt tests for completeness and referential integrity, anomaly alerts on key metrics, and a documented metric catalog with change control. Each cycle produced an immutable snapshot in Snowflake. Finance and Strategy approvals were required before publication, and the workflow logged approver identity, data version, and template version. Row-level security and role-based permissions protected pre-release data.

How did you roll this out without disruption?
We ran the automated pipeline in parallel with the legacy process for a cycle, reconciled differences, and tuned definitions. Once stakeholders were comfortable, we enabled the approval gate and made the automated deck the primary path, keeping a contingency route in reserve. Existing systems remained; the orchestration layered governance and repeatability on top.

How were currency and calendar cut-offs handled?
Exchange rates and cut-off logic were encoded in dbt models with clear ownership. The snapshot process applied those rules consistently across functions and regions. If a late regional update required an exception, it entered a human-in-the-loop review and, if approved, a new snapshot was generated and routed through the same sign-off steps.

Did this replace our existing BI or board portal?
No. Power BI remained the visualization layer, and Nasdaq Boardvantage remained the publication and distribution platform. The change was the addition of a governed pipeline, a shared metric catalog, and an approval workflow that ensured the deck was always built from a controlled, auditable snapshot.