NoteDue to the sensitive nature of this work, some images are blurred or omitted — only publicly available visuals are shown.
What a cash cost curve is — and why it mattered
A cash cost curve ranks every producer of a chemical by what it costs them to make it, plotted against capacity. Where your assets sit on that curve — and where the demand line crosses it — tells you what the market price should be and how much room you have to negotiate. At the Rotterdam workshop, marketing managers voted the cash cost curve the single highest-value thing the tool could give them.
The first version delivered that view: curves built from consultant models, filterable by product, region, and supply scenario, with demand, marginal cost, and market price lines overlaid — and Shell's own assets highlightable at a glance.
The problem: a curve you can't question
A curve you can only look at is a black box — and "no black box" was the loudest requirement users gave us. Analysts didn't just want to see the consultant's model; they needed to interrogate it, adjust its assumptions, and maintain their own view of the world alongside it. That desire for self-sufficiency — "creating our own view" rather than relying on third parties — had been on the wall since day one of the workshop.
"Need to interrogate inputs and assumptions. Retaining business logic. NO BLACK BOX."Workshop risk list, Rotterdam
The design: every assumption, inspectable and editable
CCC 2.0 turns the curve from a picture into a model. Behind every bar on the chart is an asset — a real plant — and users can open it: plant info and notes, the yield table, capacity and selectivity, fixed costs, pricing, and product transport costs, each a tab in an edit-asset workspace.
The hardest design problem was the yield table: a deep matrix of feedstocks against coproducts, utilities, and extraction stages — spreadsheet-grade density that had to stay legible and safely editable in a web tool. The answer was aggressive progressive disclosure: collapsible row groups for coproduct families and extraction stages, one clearly-focused editable cell at a time, and explicit save/cancel affordances so exploration never silently became a commitment. The Excel tool this product replaced had failed precisely because power came without integrity — CCC 2.0 had to offer the same power with it.
Finding the right asset had its own complexity: filtering cascades from product to region to technology to producer, and the available technologies change with the product chosen — a steam cracker matters for ethylene, not for MEG. The data panel bakes that logic in, so users can't build a nonsensical view.
Outcome
CCC 2.0 moved the cash cost curve from a consultant's deliverable to a living model the business owns — assumptions visible, adjustable, and annotated with the team's own notes and news. It closed the loop on the product vision from the first workshop: a tool that can be trusted because it's transparent, across all roles and regions.