From Zero to One: A Four-Stage Path to a Hedging System

The previous five articles covered process, strategy, and the pain points of three kinds of company: traders, manufacturers and policy-mature organisations. This one draws them into a single implementation path — answering one concrete question: in what order should a company do what?

Get the order wrong and the investment is wasted. That is not alarmist: we have seen companies buy systems before the policy was written, roll out to a full product range before running a pilot, and declare "hedging does not work" without ever performing a single attribution.

Each step costs real money, and the correctness of the next step depends on the output of the last. So the sequence itself is the most important decision.

Stage 1 · Foundation (4–8 weeks): turning "I don't know" into "I know"

This stage pursues no automation and no polished reporting. It does one thing: work out what exposure the company currently carries, and write down the most basic rules.

The four things you must complete

1. Exposure inventory. Pull together every contract in hand, every order in transit, every inventory position and every open futures position; rebuild net exposure along three dimensions — product, customer and contract. Error is acceptable, blanks are not. The deliverable is an exposure list, and it should be a living document that can be challenged and updated, not a one-off inventory report.
2. Policy framework. It does not need to be complex, but it has to settle four things: who may decide to hedge, within what limit, with which instrument, and when escalation is mandatory. If those four cannot be written down, everything after them is empty talk.
3. Instrument fit. Answer product by product: can futures solve this price risk, and to what extent? For the remainder, basis pricing or options? Do you need a product that is monitored rather than hedged? A conclusion of "we do not hedge this product for now" is professional judgement, not failure.
4. Small-scale pilot. Choose one product, one or two customers and a controlled size, and run the whole process for real, including exception handling, review and correction. The pilot is not there to make money; it is there to prove the process works end to end.

The mistakes this stage makes most often

The most common is skipping the inventory and going straight to a ratio. A ratio is an output of the inventory, not a starting point. Setting a ratio before exposure is calculated and then executing it mechanically is not risk management — it is a lock on risk with no dial.

The second most common error is writing the policy far too thick in one pass. A thirty-page, hundred-clause hedging policy gets signed and filed, and in practice people carry on doing what they always did. The first version should be thin enough to be read end to end within a month and explained to everyone.

Stage 2 · Embed (3–6 months): handing the rules to the system

Once the foundation is done, the company's need turns naturally into a specific question: can these rules stop depending on people's memory?

The core of this stage is turning manual process into system process, focused on four things:

• Process into the system. Pricing requests, approvals, orders, fills, attribution and settlement should connect end to end inside the system, instead of being scattered across email, chat groups and a few spreadsheets.
• Rules parameterised. Limits, tolerances, approval thresholds and pricing rules move from words in a document to configuration in the system. This is the dividing line between mechanism and policy — a parameter can be adjusted, traced and enforced; a sentence can only be read.
• Automate the routine. Pricing requests, limit checks, fill attribution and report generation: once these four are automated, exposure can be stated on one definition and in real time.
• Monitoring dashboard live. It must show net exposure, not positions. This is the step most easily done wrong in the whole system, and the one that most needs to be settled first.

For reference, industry practice sets the following targets for a systemised setup: attribution compressed from 8–12 minutes to under 1 minute per trade, error rate below 0.3%, historical retrieval in 3 seconds per 10,000 records, routine night-session requests handled automatically above 90%, and manual staffing time down by 60% or more. The same practice sets the targets for contract rollover: about 20 minutes per batch compressed to under 3 minutes per batch, manual checking time down by more than 70%, and the mis-operation rate held below 0.5%. The significance is not speed. It is that for the first time the company holds a single-definition exposure dataset that can be inspected at any moment.

Stage 3 · Extend (6–12 months): from point hedging to a strategy system

Only when the basic process is stable does the company earn the right to talk about "strategy".

Three things happen here:

Extend the product range. Copy the logic that already works to other products. But note the condition: the previous product must already run without manual intervention. If the last product still depends on someone watching exposure, extending only multiplies the staffing problem.

Layer the strategies. Static, rolling and dynamic are not a menu of one; they are used in layers according to the nature of the exposure: signed and priced exposure uses static, continuous cross-period exposure uses rolling, and exposure whose quantity itself swings uses dynamic. Each layer needs its own rule, its own assessment and its own review definition.

Widen the toolkit. Beyond futures, introduce basis pricing, options and collar structures to solve the problem of giving up too much favourable movement while hedging. This step requires clean exposure data; without it, an option is just another form of gambling.

The completion test for this stage is a number that can be checked: hedge effectiveness can be quantified — not to prove that hedging works, but to know where, in which product and at which point, hedging is working and where it is not.

Stage 4 · Operate: turning mechanism into capability

By this stage the company no longer needs to ask whether to hedge. The question is a different one: how do we keep this mechanism producing value?

Concretely, four long-run activities:

• Research supports decisions. The point of monthly commodity research is not market commentary; it is a direct operating instruction — how much to hedge now, and with which instrument.
• Stress testing becomes routine. Run reverse scenarios against current positions and turn "what if price falls 15%" into a specific date and a specific amount, so finance can prepare funding in advance. This is the most effective single defence against being forced to liquidate.
• Assessment and review. Show hedging contribution on its own line, separate from operating results, and review hedge ratio reasonableness periodically rather than once a year.
• Audit trail. Any single operation can be fully reconstructed six months later — who did it, on what rule, and how exceptions were handled. Get this right and internal and external audit shift from "keep asking for material" to "accept it directly".

There is only one metric to watch at this stage: whether the gross margin volatility range has narrowed. Not the number of hedges, not the hedge ratio, and not futures profit — those are all process measures.

The four stages at a glance

If you can only afford one thing

For almost every company the first unit of investment should go to the exposure inventory in Stage 1. The reason is direct:

Without exposure clarity, every later step is guesswork. Write policy on a guess and the policy detaches from reality. Buy a system on a guess and the system faithfully amplifies the error. Hedge on a guess and the position size is a bet either way.

And the exposure inventory is the only one of the four that is cheapest, fastest to show value, and independently valuable: even if you do nothing else afterwards, you now know how much risk you carry, where it concentrates, and how much you would lose in the worst case. That alone repays the cost.

An action you can start today

Do not wait for a proposal, a budget or a system.

The first step available today is this: pull the contract ledger and the futures positions for one product and one customer onto the same sheet, reconcile them line by line, and explain every difference.

If you can explain 100% of them, you have a starting point you can scale. If you cannot, the unexplained differences are the risk you are carrying today without knowing it. That sheet is usually the beginning of the whole system.

Sources and notes

The four-stage path and the deliverables designed for each stage are RMC's standard implementation methodology for corporate commodity risk management system projects. The practice targets cited here — attribution in under 1 minute per trade, error rate at or below 0.3%, historical retrieval in 3 seconds per 10,000 records, automatic handling of 90% or more of routine night-session requests, and manual staffing time down by 60% or more — together with the contract rollover targets of about 20 minutes per batch compressed to under 3 minutes per batch, manual checking time down by more than 70%, and the mis-operation rate held below 0.5% — come from RMC's actual operating data across delivered projects. This article uses no specific client names or data. Nothing in this article constitutes a recommendation or evaluation of any specific software product.