// Case study / BI enablement
The reporting layer survived the people who built it.
A BigQuery, Dataform, and Looker Studio reporting build replaced recurring manual report pulls, then a structured handoff made the system maintainable by the client team.
What happened
A manual reporting cycle became a maintained analytics system.
A state-regulated retail platform moved from two analysts spending the equivalent of nearly a combined work-week per month manually pulling data to a unified BigQuery, Dataform, and Looker Studio stack with structured enablement behind it.
The visible metric was 38 analyst hours recovered per month. The more important result was durability: when the client team turned over, new analysts required zero ramp-up time on the data infrastructure.
Situation
The data existed, but ownership was trapped in manual work.
Web, Braze, Google Search Console, and additional third-party data lived in disconnected sources with no shared reporting infrastructure. A monthly reporting cycle required manual extraction, reconciliation, and formatting. Domain specialists had no clean self-serve path into their own data.
Task
Build the system and the conditions for adoption.
The work was not only to unify the data layer. It was to design a handoff that would survive team turnover: clear documentation, repo access, branching and review practices, data dictionaries, ERD summaries, and training sessions built for the analysts who would inherit the system.
SELECT hours_recovered, handoff_sessions, ramp_up_days
FROM engagement.adoption_handoff
WHERE client = 'sslp_retail';38hrecovered / month
3handoff sessions
Day 1incoming analysts operational
GTM container structure, GA4 event taxonomy, Looker Studio architecture, and BigQuery web pipeline from end to end.
BigQuery models, Dataform schedules, Looker Studio surfaces, and line-by-line walkthrough of production SQL.
Braze dashboard orientation, Python automation walkthrough, API key rotation, and operational checklist.
Confluence knowledge base, ERD summary, data dictionary, repo access, branching strategy, and PR review process.
Action
The reporting layer was redesigned around self-service.
Web, Search Console, Braze, and additional third-party reporting data were unified into BigQuery via scheduled Dataform pipelines; Looker Studio dashboards gave domain specialists cross-channel visibility; and a Python automation against the Braze REST API removed a persistent manual reporting task.
The Braze REST API automation handled a recurring demand-marketing reporting task: dynamic campaign-name matching and explicit exclusion rules replaced manual Braze UI extraction, writing send volume, open rate, click rate, unsubscribe rate, and revenue attribution directly into the team’s Google Sheet template.
Documentation was written for a client analyst audience, not an agency audience. The goal was operational clarity: what runs, where it lives, how to validate it, and what to do when ownership changes.
Outcome
The client team could work from domain expertise on day one.
- 38 analyst hours recovered per month, redirected from data pulling to analysis.
- Three structured handoff sessions across web, reporting, and lifecycle layers.
- Zero data-infrastructure ramp-up when the full client team rolled off in June 2025 — incoming analysts were operational on day one.
- System has been running independently since March 2026, when the build team transitioned to a new engagement. No escalations back.
- Braze reporting automation added, saving roughly two hours per month and reducing monthly campaign reporting to zero Braze UI clicks.
- BigQuery MCP added as a follow-on enablement layer, helping non-technical stakeholders explore governed BigQuery data without waiting for recurring analyst pulls.
Design lesson
A handoff is part of the architecture.
Documentation, repo practices, and training design are not administrative leftovers. They are the adoption layer of the analytics system. Without them, the pipeline works only while its builder is nearby.