Price decisions
in context
Predicta: supporting fuel pricing decisions across stations and fuel types.
- Design lead
- Lorenzo Rogante
- Design scope
- Head of Design / Design Lead
- Collaboration
- Management and development team
- Designed for
- Pricing analysts and network operators
Pricing decisions combine network priorities with local conditions. The design connects the overview to station detail while preserving the selected station and fuel, so the investigation can deepen without repeating the initial selection.
or explore the workspace
Key decisions
Overview into investigation
Network indicators lead into the stations and fuels behind the totals.
Comparable exposure
A station-by-fuel matrix combines values and labels with colour to support comparison.
Visible assumptions
Expected impact, missing inputs and price limits remain explicit before a proposal can be assessed.
Volume and revenue at risk
Exposure = daily volume × churn probability; revenue exposure also includes the current €/L price. Illustrative prioritisation, not a validated forecast.
Make the overview a starting point
Volume, margin and recent changes sit together in the dashboard, with a route into the stations behind the totals. Summary indicators support choosing an investigation rather than becoming an isolated report.
A price worth investigating
Current → Proposed · €/L · A synthetic estimate
Compare stations and fuels on a shared basis
The station-by-fuel matrix organises exposure across the network. Volume and revenue views use distinct units, with values and labels supporting the colour scale. Each cell leads into a specific station and fuel.
A price worth investigating
Current → Proposed · €/L · A synthetic estimate
Keep the selection as detail opens
Current price, proposed price and difference appear together. Station and fuel remain selected through the transition, preserving the context of the comparison.
A price worth investigating
Current → Proposed · €/L · A synthetic estimate
Expose the trade-off and missing inputs
A price increase can affect sales volume as well as margin per litre. The comparison shows both, including a synthetic unfavourable scenario and an example where missing cost data prevents an estimate.
Read the conditions
Current → Proposed · €/L · A synthetic estimate
Keep the market comparison available
Nearby competitor prices and recent movements sit alongside the proposal. The operator can inspect the local context before assessing a response. Locations and market records in this reconstruction are invented.
Read the conditions
Current → Proposed · €/L · A synthetic estimate
Make the operating limits explicit
Price limits, discounts, market references and approval settings remain close to the proposal. Conditions can be adjusted locally in this reconstruction; they do not update a real station.
Read the conditions
Manual approval · Price limits · Market reference
A proposal stays connected to its assumptions
This case study presents professional product design work. Interfaces have been recreated for the portfolio using synthetic records and generic names. Interactive sequences illustrate design decisions rather than record the original application. The figures are examples, not measured project outcomes.
A decision workspace for fuel pricing
Fuel pricing involves a network of local decisions. A signal at one station needs to be understood in relation to the fuel involved, the surrounding market and the potential effect on volume and margin.
My work on Predicta, presented here as Pricing Intelligence, focused on giving that information a coherent interface: a way to scan the network, investigate a specific station and evaluate a proposed price action.
The challenge
The experience brings together station performance, fuel-level comparisons, market signals and AI-supported pricing recommendations. These layers need to remain connected as the user moves between the overall network and an individual pricing situation.
The design challenge was to make priorities understandable while preserving the context needed to evaluate them. A recommendation becomes useful when the user can see the current position, the relevant market information and the estimated consequences of a potential action.
My role
As Head of Design / Design Lead at BackwellTech, I helped shape the product direction and defined its user experience and visual language. My work covered information architecture, interaction flows, interface design, reusable components, prototypes and collaboration with developers.
The design included network-level views, a station-by-fuel matrix, station details, pricing actions, performance impact and market analysis. I connected these areas through a shared interface structure so the same pricing situation could be examined at different levels.
A representative decision path
The user reviews a pricing signal for a station and a specific fuel, then examines the station’s current position and the relevant market context. The proposed price can be considered alongside the potential effect on volume and margin.
This is the decision path illustrated by the case study: identify an item needing attention, understand its context and evaluate the proposed action. The flow is designed to preserve station and fuel context throughout the investigation.
Three design decisions
1. Make the network scannable by station and fuel
I used a station-by-fuel matrix to organise the relationship between locations and the products being priced. This gives the network a consistent structure and provides a route into the specific situation that needs investigation.
The visual model brings together measures such as volume or margin at risk and indicators of customer-loss risk. These signals need clear meaning so users can distinguish relative importance from the type or level of risk.
The intention is to support prioritisation across the network while keeping station and fuel context visible. The interface also distinguishes items requiring attention, waiting items and recent activity.
2. Connect network signals to local market context
I connected the overview with station-level detail and market analysis. A network signal provides a starting point; investigating the individual station provides the context for assessing a pricing opportunity.
The detail experience brings together pricing actions, performance information and relevant settings. Market context helps the user evaluate the station’s position in relation to the surrounding competitive environment.
This supports a continuous path between a broad network view and a focused local analysis, so the user can understand why a signal matters in that particular location.
3. Keep the proposed action and its expected impact connected
I structured the pricing experience to connect the current price, the AI-supported proposal and the estimated impact. Each has a different meaning in the decision: the current position describes the starting point, the proposal describes a possible action, and the projection describes its anticipated consequences.
The interface needs to preserve those distinctions when presenting volume, margin and performance information. A projected opportunity is an input to a decision; assessing the result after an action requires observed data.
Connecting the action to its expected impact gives the proposal an explicit rationale. Follow-up performance information can then support the assessment of what happened after a price change.
Design outcome
The work established a UX and visual foundation linking network prioritisation, station detail, market analysis and the evaluation of pricing proposals. It produced an interface structure for working with complex pricing information across different levels of detail.
The design outcome is a shared structure for interpreting priorities and evaluating a pricing situation, from the overall network down to the individual fuel.
What I would evaluate next
I would evaluate whether users can identify a station and fuel requiring attention, understand the meaning of the visual risk indicators, and assess a proposal using the relevant market context. I would also check whether current performance and projected impact remain clearly distinguishable.