A unified platform giving Shell's senior leaders real-time visibility into the operational and financial health of 100+ terminals worldwide — turning raw Databricks data into decisions about which terminals to invest in, optimise, or shut down.
Shell's global terminal and LNG business generates enormous volumes of operational and financial data — but in 2025, that data was scattered across siloed tools, manual reports, and fragmented systems. Senior leaders had no single view of what was actually happening across the portfolio.
In parallel, Shell was investing heavily in Databricks to centralise and stream data from terminals worldwide — covering everything from tank inventory and throughput to OPEX, HSSE events, and P&L. The infrastructure existed. What didn't exist was a way for leaders to act on it.
The question put to our team: what do we do with all this data, and how do we turn it into decisions?
Before — 12+ fragmented tools, manual packs, and no single source of truth for terminal performance
I was embedded in a cross-functional team of Business Analysts, Product Owners, and developers as the lead designer. My job was to work with this team to determine which operational and financial data capabilities were most valuable, shape them into a coherent product, and design the experience that senior leaders would actually use.
This wasn't a straightforward feature build. It required making sense of a sprawling data landscape, aligning stakeholders across commercial, operations, HSSE, and finance, and translating complex capability discussions into clear product decisions.
The first phase was understanding what data was actually available and what decisions it could support. Working closely with BAs and the Databricks team, we mapped the data capabilities coming through the pipeline:
With the BAs, I facilitated workshops to prioritise which capabilities were most decision-critical and which personas needed them. The output was a capability-to-persona matrix that became the backbone of the product.
Capability-to-persona matrix — mapping Databricks data streams to decision-critical use cases per role
The platform needed to serve radically different users — from terminal operators reviewing live asset status, to traders managing customer positions, to VPs making portfolio-level investment calls. A one-size-fits-all dashboard would serve nobody well.
The solution was a role-based architecture: a single platform with purpose-built landing pages and navigation paths for each persona, all drawing from the same Databricks data layer. Every role gets the right view of the same source of truth.
VP Operations view — portfolio health across 100+ terminals, ranked by underperformance signal
The Terminal Performance Center gives senior leaders — VPs of Operations, VPs of Finance, and regional executives — a single view of portfolio health. Crucially, it surfaces the data needed to answer the hardest business questions:
OPEX drill-down — unit cost vs. benchmark with variance trends, by terminal and region
P&L view — revenue, direct costs, and notional profit at terminal level with divestment signal indicators
HSSE dashboard — incident rates, near-miss counts, and audit completion surfaced alongside operational data
Savings tracker — FIT4 and Terminal Transformation delivery vs. target, with variance and risk flags
The Terminal Performance Center replaced 12+ fragmented tools with a single role-based platform. By working with the BA and engineering team to define which Databricks capabilities to expose first, we prioritised the highest-value data for senior decision-makers while building a foundation that scales to new terminals, new personas, and new data sources over time.
The platform positions Shell's leadership to move from lagging, manual reporting to near real-time, data-led decisions — including the ability to identify terminals that are not commercially viable and build the evidence case for shutting them down.
Next case study