NoteDue to the sensitive nature of this work, some images are blurred or omitted — only publicly available visuals are shown.
The challenge
Shell's Chemicals Commercial Business wrote contracts annually using what they themselves called an "art-based" approach — gut feel, personal spreadsheets, and institutional knowledge. To grow revenue they wanted data-driven pricing decisions, with the ability to track actuals alongside market insights.
They'd tried before. An external vendor had built a pricing tool in Excel that never met the business's needs, was extremely challenging to use, and had been effectively abandoned. Stakeholders were exhausted from talking about it. Trust — in tools, and in data — was the real starting deficit.
Kickoff: a design sprint in Rotterdam
The Pricing Excellence Tool began with a four-day Launchpad workshop in Rotterdam that I co-led with a service design lead, service designer, and product manager, facilitating among 15+ international participants from across the chemicals business.
Project background, stakeholder mapping, ecosystem map, and "Dear John" letters to past failed tools — structured reflection on what went wrong.
Built personas for Commercial Analysts, Marketing Managers, and Sales/Product Managers; mapped use cases and voted on the most frustrating and most valuable steps.
Co-created problem statements and a north-star product vision; wrote user stories by product line.
Participants sketched the tool themselves; we converged on a layout, prioritized the feature backlog, and named the risks.
The insight that reframed everything
The problem arrived at the Agile Hub framed as a technology problem: "the tool needs to be improved." The workshop revealed the technology issues were symptoms. The real problems: data lived in too many places, and nobody fully trusted it.
Commercial Analysts — who wanted to be strategists — were spending their days acting as human data libraries for everyone else's requests. Gathering pricing data was voted the single most frustrating step by every user group. And the group was emphatic about trust: no black box. Users needed to interrogate inputs, see source formulas, and keep business logic transparent.
So the team aligned on a foundation-first strategy — build a central data repository (a data lake) before the analytics — captured in a shared vision:
"An easy to use tool that can be trusted and is transparent across all roles and regions so that users have the power to make informed pricing decisions to maximize margins."Product vision, co-created at the Rotterdam workshop
End-to-end design
From there I owned design across the product's life: research touchpoints with users (scheduled with the week's notice they asked for), journey maps, wireframes, prototypes, and UI specifications for the development team — working in sprints alongside the PM, tech lead, and 12 developers, embedded with a product owner and subject-matter experts from the business.
The dashboard wireframe sketched with users in Rotterdam became the seed of the product: a data panel for products, regions, sources, currencies and units on the left; charts with comments and export on the right. Sprint by sprint that grew into the real thing — a Price Viewer where users plug their actual contract formulas in and run them against industry markers to see if their contracts beat the market; Supply & Demand modules with outage tracking to spot tightening markets; manual price forecasting; and personalized dashboards.
Two design principles carried through every sprint, straight from the workshop's risk list: transparency — every chart can show its source formula and data lineage, because trust was the product's real currency — and consistency across roles and regions, so an analyst in Houston and a marketing manager in Singapore read the same numbers the same way.
Designing the system
Shipping one screen at a time for two and a half years only works if the parts agree with each other. I built and maintained the product's design system — a UI kit (first in Sketch, later Figma), a color palette tuned for dense data visualization, and a consistent architecture every module inherits: a dark data panel on the left for products, regions, scenarios and units; a light chart canvas on the right with export, chart options, and comments.
That skeleton let the product grow to five modules — Supply & Demand, Price Viewer, Cash Cost Curve, Value in Use, and user Profiles — without users relearning anything. It also carried the collaboration layer the workshop asked for: comments and replies on outages, super-user roles for maintaining shared data, notifications, and the ability to save, organize, and share analyses from a personal account — so an analysis stops being one person's spreadsheet and becomes something the whole team can pick up.
Outcome
The tool's first release nearly tripled weekly active users — from 28 to 80 — with parallel gains in shared analyses and time in tool. Development continued sprint after sprint for two and a half years, expanding to more product lines and integrating with a sister deal-analysis tool. It's now used across Shell's global chemicals business.
The result that mattered most to me was the one from the workshop's closing circle:
"I feel like we had to go through this process to get to this point. Now this feels like one team. It feels like our solution."Workshop participant, Chemicals Commercial Business