Overview
Our client is a leading manufacturer of agricultural and industrial equipment operating across international markets, with a B2B sales model distributed through an authorised dealer and reseller network. The client's website serves as a primary discovery and research channel for prospects and distributors, where meaningful conversion actions centre on product brochure and playbook downloads rather than direct online transactions. Despite access to Google Analytics 4, the client had no reliable mechanism to trust or act on its own web data, with sampled GA4 reporting, undefined conversion events, and the absence of a structured data pipeline meaning internal stakeholders were making website and content decisions without confidence in the underlying numbers. NeenOpal partnered with the client to rebuild the analytics data foundation from the ground up, engineering a modern pipeline from GA4 through BigQuery into a multi-layer Snowflake data warehouse, and delivering a suite of Power BI dashboards that provided trusted, unsampled website intelligence for the first time.
90%
Reduction in Data Query Time
10x
Faster Report Generation
Customer Challenges
Decisions Built on Data Nobody Trusted
GA4 Data Was Sampled, Incomplete, and Untrustworthy
Google Analytics 4's native reporting applies data sampling to large datasets, which introduced systematic discrepancies between what stakeholders saw in GA4 dashboards and what the underlying data actually showed. The client's team had no visibility into the extent of this sampling or how significantly it was distorting reported metrics, causing persistent confusion around user counts, session behaviour, and engagement trends.
No Reliable Dashboard Infrastructure for Internal Stakeholders
The organisation relied on a Looker-based setup that lacked the depth and integration required for the client's evolving analytics needs. There was no centralised, structured environment where marketing, content, and leadership teams could view consistent, up-to-date performance data, leaving questions about underperforming URLs, drop-off behaviour, and engagement-driving content unanswered.
Conversion Tracking and Consent Management Were Undefined
The client had not established a clear framework for what constituted a meaningful conversion on a B2B manufacturing website, where brochure and playbook downloads function as lead indicators rather than revenue events. Consent management rules were also insufficiently defined, raising concerns about the legitimacy of tracked leads and whether compliance requirements were being met.
Complex GA4 Data Structure Blocked Downstream Analytics
GA4 exports data to BigQuery in a highly nested, event-level schema that requires significant transformation before it can be consumed by BI tools. The raw format stores user interactions as arrays of event parameters, which cannot be queried directly in standard SQL or loaded into visualisation platforms without prior flattening, leaving the data lake populated but analytically inaccessible.
Solutions
A Modern Analytics Stack Built End to End
01.
Data Lake and Warehouse Architecture
BigQuery was configured as the data lake to receive raw event-level exports directly from GA4, and a Snowflake data warehouse was engineered with a three-layer architecture covering a raw bronze layer, a cleansed and normalised silver layer, and a fully transformed analytics-ready gold layer. This layered approach ensured traceability from source to dashboard while protecting the integrity of downstream reporting.
02.
ETL Pipeline Engineering for GA4's Nested Data
Custom ETL pipelines were built from the ground up to handle GA4's complex nested data structure, unnesting event arrays, standardising column naming and formatting, and producing clean tabular output in Snowflake's gold layer. By routing all reporting through Snowflake rather than querying GA4 directly, the solution eliminated the sampling that had previously distorted the client's reported metrics.
03.
Power BI Dashboard Suite
A comprehensive suite of Power BI dashboards was developed and published via service accounts on Power BI Server, with a daily automated refresh cycle to ensure stakeholders always accessed current data. The suite covered website performance, user behaviour, geographic distribution, a CEO-level summary view, and old versus new URL comparisons to support migration and restructuring decisions.
04.
Conversion Framework for B2B Lead Intelligence
Given that the client's website does not facilitate direct sales, NeenOpal worked with the team to define a B2B-appropriate conversion framework in which brochure and playbook downloads served as the primary lead indicators. These conversion events were formally configured in GA4, correctly tagged, and carried through the full pipeline into dedicated dashboard views aligned with the client's distributor-driven sales process.
05.
Data Governance and Role-Level Security
Role-level security was implemented across all Power BI dashboards, ensuring each stakeholder group accessed only the data slices appropriate to their function. Snowflake access was restricted to a small, named group of administrators across NeenOpal and the client organisation, and consent management tracking was formalised within the GA4 configuration to address compliance and lead legitimacy concerns.
Stop making decisions on data you don't trust
Get in touchServices
Benefits
Fully Trusted, Unsampled Web Analytics for the First Time
By routing all reporting through the Snowflake warehouse rather than querying GA4 directly, the client received website performance data that reflected the complete, unsampled dataset. Metrics previously distorted by GA4's sampling engine, including user counts, engagement rates, and content interaction figures, were now accurate and consistent across every dashboard view.
Actionable Insight Into User Behaviour and Content Performance
The Power BI dashboards gave marketing and content teams direct visibility into where users engaged, where they dropped off, which URLs drove the most value, and what content was being downloaded across geographies. Scroll depth tracking and URL performance comparisons enabled evidence-based decisions about website structure and content prioritisation.
A Scalable Data Architecture Built for Long-Term Growth
The three-tier Snowflake architecture provided a foundation designed to accommodate growing data volumes and evolving analytics requirements without structural rework. New data sources can be introduced at the bronze layer and propagated through the transformation pipeline without disrupting existing gold-layer tables or downstream dashboards.
Enterprise-Grade Governance Across All Stakeholder Levels
The role-level security implementation ensured that sensitive website performance data was accessible only to authorised users, with a clear, administrator-controlled permission structure maintained by the client's own team. The CEO-level dashboard provided leadership with a governed, accurate summary view, while granular operational views were scoped appropriately to marketing and content functions.
A Defined B2B Conversion Framework Supporting Lead Intelligence
By formalising brochure and playbook downloads as the client's primary conversion events within both GA4 and the downstream dashboard layer, the engagement gave the marketing team a principled and consistent measure of website effectiveness. Lead indicator tracking allowed the client to assess which content drove the highest-quality engagement and which product areas were generating the most prospect interest.
Conclusion
For a global B2B manufacturer whose website is its most scalable channel for prospect and distributor engagement, the inability to trust web analytics data was not a reporting inconvenience but a strategic liability. By engineering a complete data pipeline from GA4 through BigQuery into a governed Snowflake warehouse, and delivering a Power BI dashboard suite with role-appropriate access for every stakeholder level, NeenOpal gave the client the analytics foundation it needed to make confident, data-driven decisions about its digital presence. The solution resolved years of data distortion introduced by GA4 sampling, established a clear B2B conversion framework aligned to how the business actually generates leads, and put in place a scalable architecture that will continue to deliver value as the organisation grows its analytics maturity.
FAQ
Frequently Asked Questions
Why was a separate data warehouse needed if GA4 already provides analytics?
GA4's native reporting applies data sampling to large datasets, which means the figures shown in the GA4 interface are statistically estimated rather than based on the complete data. For an organisation making website strategy and content decisions, this introduces a systematic inaccuracy that compounds over time. By exporting GA4's raw event data to BigQuery and transforming it through a Snowflake data warehouse, NeenOpal ensured that every metric in the Power BI dashboards was derived from the full, unsampled dataset. The warehouse also decouples reporting from Google's platform, giving the client full ownership and portability of their analytics data.
What is the bronze, silver, gold architecture and why does it matter?
The three-layer data architecture is a standard pattern in modern data engineering designed to separate raw data ingestion from transformation and reporting. The bronze layer stores raw GA4 event data exactly as it arrives from BigQuery, preserving a complete and auditable source of truth. The silver layer applies cleaning, deduplication, and structural normalisation to produce consistent, well-formed tables. The gold layer contains fully transformed, analytics-ready data that feeds directly into Power BI without further processing. This layered approach means that transformations are transparent and reversible, raw data is never overwritten, and the reporting layer can be modified or extended without touching the ingestion pipeline.
How does role-level security protect sensitive analytics data?
Role-level security, implemented within Power BI, restricts what data each user or group can see when they access a dashboard, even when multiple stakeholders view the same report. The access rules are defined during dashboard development and enforced at query time, so a regional marketing manager, for example, would see only their geography's data while a leadership view shows aggregated performance across all regions. Permissions are managed through Microsoft's authentication infrastructure and administered by the client's own IT team, ensuring that the client retains full control over who can access which analytics views without depending on NeenOpal for ongoing access management.
Authors
Contact Us
We’d love to hear from you.
Lets discuss how we can transform your business with AI. Talk to our AI expert team. Lets do AI journey together.
