Redesigning how a $2 trillion industry buys ocean freight
The first index-linked contract simulator in the world, and the design work that turned a complex commercial mechanism into a product customers across the ocean freight market actually use.
Knowing the market doesn't fix the contract
- Freight was opaque: rates negotiated behind closed doors, with no market reference
- Xeneta turns 700 million crowdsourced rates across 170,000 trade lanes into a benchmark: transparent and rules-based
- But freight is still bought on annual fixed-price deals that take months to negotiate, and the market moves constantly
- The moment it moves, one side is out of pocket and everyone is back at the table: months of expensive procurement time, burned over and over
Fixed-price contracts can't hold in a moving market. So customers renegotiate, constantly.
An annual rate takes months to agree, and the market moves the day after it's signed. The moment it does, one side is out of pocket, everyone is back at the table, and the supply chain wobbles while they argue.
Index-linked contracts: tie the rate to the index, and the renegotiating stops.
The price moves with the market by rule, not by argument. The mechanism works. But it's complex, unfamiliar to most of the industry, and it rests on commercial judgement calls customers have to feel confident making.
How do we transition an industry from fixed-price contracts to index-linked contracts?
Never a pricing problem: a trust problem. What the answer should be was not obvious yet. Everything that follows, the research, Copenhagen, and the products we chose to build, comes out of this question.
Customers trusted the data. They didn't trust the mechanics.
- Discovery with customer success managers and freight forwarders
- User testing with shippers, including Walmart and Philips
- They believed the benchmark, but couldn't picture how an indexed price would behave, or defend it to their own boss
- That research became the requirements document: transparency over automation, written down before anything was designed
Why shippers were saying no to indexing
I believe your benchmark. What I can't picture is what happens to my rate the day the index jumps.
I can't walk into my CFO's office with a formula. I need to show them what it would have done to last year's budget.
The mechanism is probably fair. But if it goes wrong, it's my name on the contract, and I can't defend maths I don't understand.
Not one objection was about the pricing. Every single one was about being able to see it, explain it, and defend it. That is a design brief.
The world's largest shipping company was modelling these contracts by hand, in Excel
- They don't sell indexing with a finished contract. They walk customers through the choices. That's how trust gets built.
- Their tooling for it: a spreadsheet, precisely the gap the Simulator existed to close
- The principle: the product should feel like the sales conversation, not skip past it
How the people who sell indexing actually sell it
We never open with a finished contract. We make the choices with them: which index, how deep the discount, where the floor sits. Signing is the last step of a conversation, not the first.
We put the spreadsheet on the screen and model their own lanes, live. Watching the number move against their history is what convinces them. Not the brochure.
A customer is ready when they can explain the contract back to us. If they can't, it dies at their next internal meeting, whatever they signed.
Their sales process was the product spec: walk through the choices together, model against real history, and leave the customer able to explain it without you in the room.
The tenets of Indexing
Transparency beats simplification
Customers defending a decision to their own boss don't want a black box. They want to be able to explain it.
Feel like the sales conversation
The product walks through the choices the way Maersk does in the room. It never skips to the answer.
The user owns the decisions
The product brings the data, the history and the consequences. The judgement calls stay with the customer.
Trust must survive the handover
Simulation and operation are one system, so the trust built at the front door carries into the live contract.
Written after Copenhagen, before any screens. Every design decision that follows argues from one of these.
A free simulator as the front door, one shared model behind it
- The front door: a free, standalone simulator: backtest a contract against real market data, no commitment
- Behind it: a live contract manager to operate the contract once signed: monitoring, re-rating, applying the rules month after month
- The easy build was two products. We built one system on a shared contract model: the object a customer configures in simulation is the object the contract manager operates.
- A data-model decision, made with engineering, so trust built in simulation carries into operation
Prototyping in code, not slides
- Functional prototypes in code, not static mockups
- Customers interacted instead of imagining, which meant sharper feedback
- Engineering aligned against something concrete: technical realities surfaced early, scope stopped drifting
- Several scoping decisions, including what to defer from V1, came out of those sessions, in conversations that wouldn't have happened over a Figma file
V1 shipped deliberately simple, and taught its own lesson
- A single-lane simulator: one trade lane modelled at a time, backtested against real market data
- It shipped with Basic and Advanced modes, which split the two people who configure a contract together into different versions of the product, and broke shared walkthroughs in live sessions
- V2 replaced them with one guided flow that teaches each rule as you configure it
- The lesson: check a decision against the product's core insight, that trust is built together, not just usability logic
simulations since launch, across 660+ users at 250+ customer companies, with sustained monthly engagement across every part of the freight market it was designed for
From unfamiliar concept to award-winning product, in under two years
- The question: could design make this mechanism trustworthy enough to adopt? The usage data says it could.
- Indexing moved from side bet to one of Xeneta's strategic surfaces, helping reposition the company from a benchmarking tool toward a strategic procurement platform
Winner, Ship Technology Excellence Awards 2025
and Finalist at the UX Nordic Awards, Future Product Days
In under two years, from an unfamiliar commercial concept to an award-winning, widely adopted product.
To decide V2, I stopped sampling customers and listened to all of them
- A pipeline over Gong call transcripts, through the Anthropic API
- Claude extracts ideas and confusion, classified by theme, stance and account, linked to the exact moment in the call
- 700+ calls: anecdote turned into evidence. V2 was prioritised from that evidence, not from opinion.
- AI used not to generate screens, but to listen at scale
Customers don't think in single lanes. They buy whole portfolios.
- The question customers kept circling: how do I protect a budget when the market moves?
- V2: a multi-lane portfolio simulator: an entire contract book under one shared ruleset
- Backtested against five years of real data, including the COVID and Red Sea disruptions
- See how the book would actually have behaved, before signing anything
Guidance, not financial advice
- Xeneta had just launched the first freight futures, listed on Euronext and settled against our index, so the simulator learned to hedge
- Hedging is a downstream event: it resolves once the contract rules are configured. The interface follows the order of the decision, not the order of the features.
- A regulatory line between guidance and financial advice, so it leads with education, checks which lanes are hedgeable, and one slider dials in coverage
- The compliance boundary wasn't something to work around. It was a structure to design with.
Xeneta Indexing today: the platform the simulator grew into, with the free simulator still the front door.
Scored at both releases, with the same instrument
V1 · Basic and Advanced modes
A solid score, comfortably above the 68 industry average. But the sessions showed the ceiling: the mode split cut the two people configuring a contract together into different products.
V2 · One guided flow
The same users, after the guided flow replaced the modes and the walkthrough became shared again. Anything above 80 is an A grade.
The jump didn't come from more features. It came from removing the split that worked against the product's own core insight.
What I learned
Trust is a UX problem
In enterprise products, especially financial ones, transparency is more valuable than simplification. Customers making commercial decisions don't want a black box. They want the ability to defend the result to someone else.
The best research happens where the product isn't
Watching how a sales team actually sells indexing taught me more about how the simulator needed to feel than any usability test could have. Domain authority belongs with the user.
AI can be research infrastructure
The most valuable thing I built for V2 wasn't a screen. It was the pipeline that let me hear 700+ customer conversations at once. That's where a lot of the craft of senior design is heading.
Thank you
World first · 1,500+ simulations · 700+ calls analysed · Award winner