Your oil and gas project is ready for an approval meeting. The estimate has a firm-looking total, but the piping quantities still depend on an unsettled layout, a major equipment quotation has expired, and nobody can point to the budget for the shutdown tie-ins. Which part of that total are you actually approving?
A useful FEED cost estimate makes those gaps visible before they become commitments. This guide gives owners and investors a review framework for scope, quantities, pricing and risk. Here, the estimate means the project’s capital cost assessed around front-end engineering design (FEED), not just the engineering fee for performing FEED.
First, identify which approval you are making
Approval to start FEED and approval to proceed into engineering, procurement and construction (EPC) are different decisions. Use your organisation’s gate definitions, and state the authority being requested on the first page.
| Decision | What the review should establish |
|---|---|
| Fund the FEED stage | Whether the selected concept warrants further development, what the FEED work will cost, and which uncertainties that work must resolve. |
| Approve EPC investment after FEED | Whether the defined project, execution plan and remaining exposure support the capital commitment being requested. |
Before FEED, request a costed information-development plan: surveys, engineering studies, supplier enquiries and the decisions each will support. Show the FEED-stage budget separately from the indicative overall project investment.
After FEED, request the evidence produced by that work and identify what remains unresolved. A limited release for long-lead equipment should name its own scope and spending limit; it should not silently become approval of the whole project.
Check estimate class against deliverables, not the label
AACE Recommended Practice 18R-97 bases estimate classification on the maturity of defining deliverables. A FEED label, elapsed engineering time or design-completion percentage does not automatically establish Class 3. The practice addresses process-industry EPC estimates, including hydrocarbon processing; it is not a blanket classification guide for every exploration, production or pipeline project.
Ask the estimator to show which applicable deliverables support the claimed class and where the package falls short. Make exceptions visible by discipline. A well-developed equipment package does not resolve an uncertain site boundary or an incomplete electrical supply concept.
For the wider progression from early studies to tender, see why construction cost estimates change as information develops.
Review scope maturity and the quality of its foundations
A document can exist without being dependable enough for the decision. An issued equipment list may still carry unconfirmed duties; an approved layout may rely on a site survey that excludes the tie-in area.
CII’s FEED MATRS research assessed deliverable maturity separately from the context affecting FEED accuracy. CII subsequently integrated that approach with PDRI–Industrial in PDRI MATRS, Version 5.0, which supersedes the standalone FEED MATRS tool.
Use that distinction to organise the review, not to invent a substitute score. For each cost-driving assumption, ask: what supports it, who checked it, and what decision could still change it?
Bring engineering, operations, procurement and project controls into the same discussion. For a brownfield package, walk through access, isolation, shutdown timing, temporary works and reinstatement. Record disagreements as actions with owners and due dates. A green dashboard should not conceal a red issue that changes the investment case.
Trace the quantities behind the largest cost exposures
Start with selected high-value or high-uncertainty packages. Choose at least one equipment item and one bulk-material package, then follow each from the engineering source to the estimate line.
For each sample, request the document revision, measurement method, unit, quantity owner and reconciliation date. Identify whether the quantity is measured, model-derived, factored from a comparable facility or held as an allowance. These are different evidence bases, even when the spreadsheet displays the same number of decimal places.
Consider an illustrative pipe-rack package: a revised route may alter steelwork, foundations, cable lengths and access requirements together. Reviewing only the steel tonnage can miss the linked cost movement. Ask how the team propagates a layout change across disciplines and prevents duplicate quantities at package boundaries.
Also challenge blank or zero-value lines. Does zero mean no work is required, another package carries it, or the quantity has not been developed? Give those states different labels. An unmeasured requirement should remain visible in the gap register until its cost treatment is agreed.
This is where quantity surveying for oil and gas EPC projects connects engineering information with a reviewable commercial baseline. The test is traceability, not the thickness of the quantity report.
Make the estimate basis explain the number
AACE’s Basis of Estimate guidance, RP 34R-05, describes documentation that makes the estimate’s purpose, scope, methods, supporting information and uncertainties understandable. Treat that document as a decision aid, not an appendix assembled after the total is agreed.
Give reviewers a short route through the evidence: the requested decision, scope boundary, source register, pricing date, execution assumptions and unresolved items. Follow it with a reconciliation to the previous approved estimate.
Separate movements caused by scope changes from quantity development, rate updates, schedule changes and corrected omissions. Otherwise, a lower total may simply reflect a cost transferred elsewhere, while a higher total may represent previously hidden work becoming visible. Neither movement can be judged responsibly from the headline alone.
Separate allowances, exclusions and risk provision
Agree the terminology used in the approval pack. For this review, an allowance means an included amount for a described requirement whose detail remains incomplete. An exclusion means a cost is outside this estimate. Contingency addresses uncertainty and risk within the stated basis; it should not be used to obscure known missing scope.
For every material allowance, name the requirement, calculation basis and evidence that will replace it. For every exclusion, identify whether it belongs in another funded package, remains an owner obligation or is outside the agreed project scope.
Build a bridge from the EPC estimate to the total investment request. Check how owner engineering, commissioning support, initial spares, temporary facilities and interfaces are treated where applicable. Do not assume an excluded item disappears from the owner’s funding need.
Connect prices to the purchasing schedule
A dated quotation is evidence of a price under stated conditions, not a promise that the same price will be available at award. Review quotation validity, specification alignment, delivery assumptions and the purchasing date in the schedule.
For imported equipment, distinguish the supplier’s pricing currency from the reporting currency. Document the exchange-rate basis and test a different rate against the exposed payments. Have treasury confirm any assumed protection; do not describe an unconfirmed hedge as secured.
Keep escalation separate from currency movement so reviewers can see what each assumption changes. Check whether quoted prices already include future price adjustments before adding another uplift.
Then test the market evidence package by package. If a critical supplier slot is unavailable at the planned date, revisit both delivery and cost. A project-specific scenario is more informative than applying one unsupported market factor to every line.
Ask what the risk range means
AACE 18R-97 cautions that scope maturity alone does not determine accuracy; the range should come from analysis of the specific project’s risks. Its typical class ranges are not a guarantee for an individual project. See the AACE classification guidance.
Ask the review team to explain the base estimate, included risk provision and resulting funding request separately. If a probability-based figure is presented, state the confidence basis, model assumptions and risks omitted from the analysis. A management target is not automatically a modelled confidence level.
Challenge linked exposures: a late shutdown can change field duration, supervision, equipment hire and purchasing dates together. Record the response to each material risk and who owns it. Do not select a contingency percentage simply because it makes the project fit the available budget.
Leave the gate with a recorded decision
Use this proposed owner checklist alongside your organisation’s approval requirements:
- Authority: Is this FEED funding, a limited commitment or the EPC investment decision?
- Scope: Are boundaries, interfaces and unresolved design choices explicit?
- Evidence: Are the claimed estimate class, major quantities and rates traceable?
- Completeness: Can reviewers reconcile allowances, exclusions and owner costs to the total request?
- Exposure: Do schedule, market, currency and risk assumptions describe the same execution plan?
- Governance: Who accepts the remaining uncertainty, closes each action and authorises later changes?
Record approval, conditional approval or deferral with a clear rationale. Conditions need deadlines and release controls. Preserve the estimate revision, supporting basis and decision record together so the next gate can show what changed.
Make a condition testable: identify the package, the missing evidence, the responsible reviewer and the commitment that cannot proceed until closure. Define what would trigger another approval, such as a changed scope boundary or a revised funding requirement. This gives the project team an operating instruction instead of an unresolved comment in the meeting minutes.
For support with a defined review scope, explore GEMS estimating, quantity surveying and risk-analysis services. Bring the estimate, source documents and the decision you need to make; agree the required deliverables and limitations before the review begins.
Frequently asked questions
Is a FEED cost estimate automatically Class 3?
No. For applicable process-industry projects, AACE 18R-97 ties estimate class to the maturity of defining deliverables, not the FEED label. Ask for the supporting deliverable assessment and recorded exceptions.
What should an owner approve before starting FEED?
Approve a defined FEED scope, budget, programme and decision plan under your organisation’s governance. Keep that authority separate from approval of the overall EPC investment. State which uncertainties FEED must resolve before the next gate.
How much contingency should a FEED estimate include?
There is no universal percentage suitable for every FEED estimate. Request a project-specific risk assessment with a clear scope basis, assumptions and funding rationale. Check that known requirements are not omitted or counted twice across allowances and contingency.
