ERW CommandEPC

THE INTELLIGENT EPC LEADERSHIP PROGRAMME

The contractor that can explain its delivery.

← Programme home
30-Session Programme · 25 Minutes Per Day · Free · For leaders in Nigeria’s engineering, procurement and construction sector
WEEK 3 OF 6 · CONSTRUCTABILITY, PROCUREMENT AND SUPPLY · SESSIONS 11–15 · WORK AT YOUR OWN PACE
This week’s assessments: Pop Quiz 3 — Constructability, Procurement and Long Lead after Session 13 · Practice Exam 3 — Procurement, Fabrication and Supply after Session 15
By Session 30: you hold a signed 90-Day Intelligent EPC Programme for your own company · Every assignment marked against a published rubric · Pass mark 70
Part Two · Engineering and the Design CommitmentSESSION 11
Constructability, Configuration and Digital Twins
Session 11 of 30 · 25-Minute Module

Why This Matters

A design can meet its technical duty and still fail on site: the module will not travel the route, the lift cannot be made, the shutdown window does not exist. Chapter 10 asks you to bring execution conditions into design early enough to change the outcome, and to be disciplined about digital models. A three-dimensional model, a live operational representation and a predictive simulation are different capabilities, and the label digital twin settles none of them. This session teaches you to ask whether the design can be delivered, to state what any digital model is for, to keep it tied to the installed reality, and to value it only through the decisions it actually changes.

25-Minute Module
6:00 – 12:00
State what the digital model is for
  • Plain language first. A three-dimensional model shows geometry. A live operational representation is meant to stay current with the real asset. A predictive simulation calculates what may happen under stated conditions. Each can be sold as a digital twin; the book treats them as different capabilities.
  • Before buying, specify three things: what the proposed twin represents, how it receives updates and which decisions it supports.
  • THE DOCTRINE

    Avoid purchasing the label without accepting the underlying service.

  • UseInformation that may be essential (Chapter 10)
    Construction coordinationGeometry, revision, package identity and status
    CommissioningSystem boundaries, tests and control configuration
    Later operationsEquipment identity, maintenance data and accepted operating information
  • Declare which information is authoritative. A visual model can display a proposed location while the released drawing defines the actual instruction. If the model becomes an execution source, the project must establish the corresponding release and change process, the same discipline Session 10 applied to releasing a coherent configuration.
If nobody can say which screen the foreman should build from, the model is a picture, not a source.
18:00 – 22:00
Value the model through intervention
  • Estimate avoidable rework through actual changed decisions. A clash detected after fabrication may permit a field correction; one detected before release may avoid fabrication changes. Their value differs. Do not assign every digital flag the average cost of a major rework event.
  • Track the chain the Operations Excellence Lens asks for: the finding, the competent assessment, the accepted change, the execution and the supported cost effect. Record findings that were dismissed and findings that arrived too late. These records improve the method and prevent inflated savings, as the Operating Intelligence Loop's record of unactioned recommendations does.
  • THE DOCTRINE

    The leadership decision is whether the information model makes delivery choices more dependable at an acceptable total cost. Visual richness does not establish engineering or construction acceptance by itself.

A flag is worth what it changed, at the moment it changed it.
Key Terms

Authoritative information

The information the project declares as the instruction to act on. A visual model may display a proposed location while the released drawing defines the actual instruction.

Useful finding and material miss

The measures of a clash or constructability service: findings that changed a decision, and consequential issues it failed to detect. The count of flags is not a measure of value.

Self-Check — Before You Submit
1. A vendor offers your fictional EPC contractor a digital twin for an upcoming project. What should you require first?
  • A. A demonstration of its visual quality on a comparable plant
  • B. Agreement to make it the execution source from day one
  • C. A statement of what it represents, how it receives updates, which decisions it supports and which information is authoritative
  • D. A saving estimate based on the average cost of a major rework event per flag
C. A statement of what it represents, how it receives updates, which decisions it supports and which information is authoritative — Chapter 10 asks for a stated purpose before purchase, so the company does not buy the label without the underlying service. Visual quality is the tempting answer, but visual richness is useful only when it supports that purpose.
2. A scan of an installed pipe-rack bay matches the model. What does that establish?
  • A. An apparent geometric match for the accessible geometry the scan could see
  • B. Weld acceptance for the bay
  • C. Support loading suitability
  • D. Compliance with the approved configuration
A. An apparent geometric match for the accessible geometry the scan could see — The book's hypothetical says a geometric match does not prove weld acceptance, support loading suitability or compliance with the approved configuration. Even the geometry claim is bounded by the scan's coverage.
3. Two identical clashes are found on one project, one before drawing release and one after the spools were fabricated. How should their value be treated?
  • A. Equally, at the average cost of a major rework event
  • B. Equally, at zero, because a clash report is only information
  • C. The later one is worth more because more money was at stake
  • D. Differently: the earlier one may avoid fabrication changes, the later one may only permit a field correction, and value follows the actual changed decision
D. Differently: the earlier one may avoid fabrication changes, the later one may only permit a field correction, and value follows the actual changed decision — Chapter 10 says value depends on intervention and timing, and that their value differs. The tempting answer is the average rework cost, which the book specifically warns against because it inflates savings.

Today’s Assignment

Write Section 11 of your plan: whether your decision chain's design can be delivered, the purpose any digital model must serve, how it stays tied to reality and where its value depends on intervention.
  • Name one design condition in your decision chain that could make delivery impractical (transport, lifting, access, temporary works, erection sequence, inspection or maintenance access) and the actual site evidence your company would use to test it. Say honestly whether that review currently happens early enough to change the outcome.
  • State the purpose of any digital model your company uses or is considering: what it represents, how it receives updates, which decisions it supports and whether it or the released drawing is authoritative. If your company has no model, write the purpose a proposal would have to meet, or explain why a simpler method serves the decision.
  • Describe how installed differences reach your records, how they are reconciled with design changes and as-built information, and what your capture methods cannot see.
  • Trace one finding, real or recent, from detection to competent assessment, accepted change, execution and supported cost effect. State where its value depended on timing, and how your business records dismissed and late findings.
  • 400–900 words, your own writing. Send me the section.
YOUR 90-DAY PLAN

Section 11 — Constructability, the digital model's purpose and its value through intervention.

How this is marked — the rubric
Names a specific deliverability condition, the site evidence that tests it and an honest view of when the review happens25
States the model's purpose, update route, supported decisions and authoritative source, or justifies a simpler method25
Shows how installed reality is reconciled and names the capture methods' coverage limits25
Values findings through changed decisions and timing, records dismissed and late findings, and avoids any average-cost shortcut25
Total · pass mark 70100
Submitting session 11
Write your answer in Word or Google Docs. The first line must read Session: 11, or the submission is refused. Upload the Word file (.docx), or switch to the Google Doc tab and paste your document’s share link.
Marked against the rubric above and returned as a written letter, saved to your progress page. Do not submit until you have passed the self-check above.
Capabilities Covered
Constructability reviewDigital model purposeAuthoritative source controlAs-built reconciliationIntervention-based value
Tick it once you have submitted.Next: Session 12 →
Part Three · Procurement, Fabrication and SupplySESSION 12
Procurement Intelligence and Supplier Qualification
Session 12 of 30 · 25-Minute Module

Why This Matters

Procurement turns a design into a commitment someone else must honour. Chapter 11 warns plainly that a fast purchase against an unstable requirement can create more delay than it removes, and a cheaper offer may simply omit scope that another includes. AI can help compare requisitions, offers and evaluations, and can point to suppliers who need further review. It cannot make a requirement stable, qualify a supplier from sparse data or decide that an invoice is payable. This session teaches you to purchase what the project actually needs, match qualification to consequence, award a complete package, run a procurement advisory that names the owner of each constraint, and keep a supplier record that does not rewrite history.

25-Minute Module
6:00 – 11:00
Match qualification to consequence
  • Evaluate the supplier's ability to deliver the actual package: technical competence, capacity, quality arrangements, relevant resources, financial position, documentation and subcontract dependencies. Evidence of previous work can support investigation, but it does not automatically establish current capacity.
  • A classification model, one that sorts suppliers into groups, may identify those requiring further review. It should explain its evidence and limitations.
  • THE DOCTRINE

    Avoid opaque exclusion based on sparse data, a similarly named company or an unsupported inferred relationship.

  • Human assessment matters most where records are incomplete. Missing information can reflect a weak information interface rather than poor physical capability. The business should establish which gaps must be resolved and which proposed controls can manage them.
  • If your company is the vendor, read this from your side: a gap in your documentation can be mistaken for a gap in your yard.
A gap in a supplier's file is a question to answer, not an answer.
16:00 – 21:00
The advisory, and a record that does not rewrite history
  • Advisory field (Chapter 11)What it lets the project director judge
    Package identityWhich commitment is at stake
    Required-on-site date and current forecastThe size of the gap
    Evidence freshnessWhether the forecast rests on sufficiently current sources
    Affected milestoneWhat the gap actually threatens
    Options, cost and decision deadlineWhat can be done, at what price and by when
  • Distinguish technical approval delay, manufacturing delay, funding delay and logistics delay. They require different responses. Expediting a supplier cannot solve an unresolved specification or an unfunded deposit.
  • THE DOCTRINE

    Procurement intelligence becomes useful when the project director can identify the owner of the constraint and approve a feasible response.

  • Learn without rewriting history. Retain the award assumptions and later changes, and compare promised and actual dates on consistent milestone definitions. Performance at shipment and at accepted site receipt may differ.
  • Separate contractor-caused changes from supplier execution. A vendor repeatedly waiting for approved drawings should not be labelled late through a target that assumes those drawings were available. Conversely, a vendor's revised promise should not replace the original commitment in the performance record. AI supports this by organising evidence, not by creating a score whose meaning nobody can defend.
Ask whose constraint it is before you ask who should chase it.
Key Terms

Evidence freshness

Whether the underlying source is sufficiently current for the decision. Dashboard refresh frequency alone does not establish it.

Procurement advisory

An exception brief showing package identity, required-on-site date, current forecast, evidence freshness, affected milestone, options, cost and decision deadline, with the delay type distinguished.

Self-Check — Before You Submit
1. A supplier-screening model excludes a fictional fabrication subcontractor because its records are sparse and its name resembles a company with a poor history. What should your procurement head do?
  • A. Accept the exclusion, because the model was trained on many suppliers
  • B. Route the supplier for human assessment, establish which gaps must be resolved and which controls could manage them, and ask the model to show its evidence
  • C. Approve the supplier, since sparse data proves nothing either way
  • D. Ask the supplier to rename the company
B. Route the supplier for human assessment, establish which gaps must be resolved and which controls could manage them, and ask the model to show its evidence — Chapter 11 warns against opaque exclusion based on sparse data or a similarly named company, and says missing information can reflect a weak information interface rather than poor capability. Approving outright is the tempting overcorrection; the book asks for assessment, not a reversal of the default.
2. The procurement advisory shows a long-lead package forecast late because the specification is still unresolved. What will expediting the supplier achieve?
  • A. A recovery proportional to the expediting cost
  • B. A faster shipment, which resolves the delay
  • C. A better supplier performance score
  • D. Little or nothing on this constraint, because the owner of an unresolved specification is not the supplier
D. Little or nothing on this constraint, because the owner of an unresolved specification is not the supplier — The book says expediting a supplier cannot solve an unresolved specification or an unfunded deposit. The advisory exists to identify the owner of the constraint; the book also warns that a fast purchase against an unstable requirement can create more delay than it removes.
3. In a fictional case, a vendor delivered six weeks after its original promise, but spent four of those weeks waiting for approved drawings from the contractor. How should the performance record treat it?
  • A. Keep the original promise, record the contractor-caused waiting separately from supplier execution, and compare dates on consistent milestone definitions
  • B. Record the vendor as six weeks late
  • C. Replace the original promise with the vendor's revised date and record it on time
  • D. Delete the package from the record as unrepresentative
A. Keep the original promise, record the contractor-caused waiting separately from supplier execution, and compare dates on consistent milestone definitions — Chapter 11 says a vendor waiting for approved drawings should not be labelled late through a target that assumes they were available, and that a revised promise should not replace the original commitment. Replacing the promise is the tempting fix and rewrites history.

Today’s Assignment

Write Section 12 of your plan: the requirement your decision chain actually needs, supplier qualification matched to consequence, a complete award and the procurement advisory.
  • For one package in your decision chain, list the requisition fields (accepted scope, specification, quantity, required date, documentation, inspection, acceptance) and mark honestly which are provisional today.
  • Describe how your company qualifies suppliers for this package, what evidence it relies on, where past work is being treated as proof of current capacity and how incomplete records are handled.
  • State what your purchase order carries from the technical evaluation and accepted deviations, and how payment events are mapped to evidence. Name one place where a template could silently drop a requirement.
  • Draft one procurement advisory entry with all the fields from today's table, the delay type distinguished and the owner of the constraint named by role.
  • State how your business keeps the original promise beside revised ones and separates contractor-caused delay from supplier execution. A supplier writes it from its side of the order. 400–900 words.
YOUR 90-DAY PLAN

Section 12 — The requirement, qualification matched to consequence and the procurement advisory.

How this is marked — the rubric
Lists the requisition fields and marks provisional inputs honestly20
Matches qualification to the package's consequence and handles incomplete records without opaque exclusion25
Links the award to the technical evaluation and maps payment events to evidence25
Drafts a complete advisory entry with the delay type and constraint owner, and keeps an unrewritten supplier record30
Total · pass mark 70100
Submitting session 12
Write your answer in Word or Google Docs. The first line must read Session: 12, or the submission is refused. Upload the Word file (.docx), or switch to the Google Doc tab and paste your document’s share link.
Marked against the rubric above and returned as a written letter, saved to your progress page. Do not submit until you have passed the self-check above.
Capabilities Covered
Requirement stabilityNormalised offer comparisonConsequence-based qualificationComplete package awardProcurement advisoryFair supplier record
Tick it once you have submitted.Next: Session 13 →
Part Three · Procurement, Fabrication and SupplySESSION 13
Long Lead Equipment and Expediting
Session 13 of 30 · 25-Minute Module

Why This Matters

Long-lead equipment is where a project quietly loses its finish date. A supplier repeats a delivery date for months, and that date tells you less than evidence that drawings were accepted and the test was booked. Chapter 12 is uncomfortable in a useful way: an expedite payment can have negative expected value, a week saved on one package can save nothing at completion, and an AI summary can describe a recovery the network does not support. This session teaches you to forecast the state of work, test the economics of a response, keep prediction and response in separate records and protect technical control while recovering.

25-Minute Module
5:00 – 10:00
Design the forecast around a useful warning
  • A forecast needs an event and a decision horizon. For the book's proposed service the event is accepted site availability by the package's required date. Factory completion, shipment and arrival remain separate milestones.
  • Issue the forecast at award and at defined review intervals, using only the evidence available at each cut-off; every input needs its own availability time. A retrospective delay explanation used in an earlier forecast is leakage.
  • Start with a transparent comparator: the vendor's current accepted milestone plan plus known unresolved predecessors. With few comparable completed packages, a well-maintained milestone review and explicit scenarios may be more defensible.
  • Right-censoring, in plain words: a package still in progress has a completion time that is not zero and not yet known. Train a duration model only on the quick packages that happened to finish and it learns that packages are quick; these records need appropriate statistical treatment.
A forecast that arrives after the last useful intervention is history, however accurate.
15:00 – 20:00
The walkthrough: a recovery that recovers nothing
  • The book's walkthrough, assumed throughout. After Session 10's revision, the supplier's forecast accepted delivery moves from week ten to week twelve, an assumed vendor forecast, not a model-generated probability. Foundations (C) finish at week eight; installation (I) needs equipment (P) and C; testing and handover (T) follows I. The original finish was therefore week sixteen: eighteen less the two weeks of slippage.
  • Assumed scenario (walkthrough)Equipment acceptedITFinish
    Supplier's revised forecastWeek 12Weeks 12–15Weeks 15–18Week 18, two weeks later
    Hypothetical paid recovery offerWeek 11Weeks 11–14Weeks 14–17Week 17, one week recovered
    Same offer, client operating interface unavailable until week 12Week 11Weeks 12–15Weeks 15–18Week 18, no improvement
  • AT THE WORKFRONT · THE WORKED CASE
    • Workfront. The book's own case. Installation must follow P, C and the client's operating interface, which R-03 links to installation work W-01 (Session 8). Week-eleven equipment would wait for a week-twelve interface.
    IN THE COMMITMENT · THE PARALLEL
    • Commitment. The AI-generated procurement summary first calls the offer one week of recovery; the integrated calculation and R-03 expose the error. The contractor does not fund a recovery the network does not support, and the project director pursues the interface decision instead.
  • Keep prediction and response in separate records: the service supplies the forecast, observations, limitations and remaining decision window; the procurement manager supplies feasible responses; the project director connects them to schedule and cost. You can then test the prediction without pretending it proves a remedy.
  • THE DOCTRINE

    A forecast's feature importance can suggest where to investigate. It does not establish that changing the feature will change delivery.

Frequent expediting can travel with late packages because managers expedite the troubled ones; the association alone does not establish the remedy.
Key Terms

Right-censored record

A package still in progress at the observation cut-off, whose completion time is not zero and not yet known. It needs appropriate statistical treatment rather than exclusion.

Common-cause scenario

A scenario in which packages sharing a port, route, vendor or client input are delayed together, which independent sampling can understate.

Self-Check — Before You Submit
1. In the book's assumed expedite illustration (option cost ₦12 million, probabilities 0.50 and 0.15), eligible consequence falls from ₦40 million to ₦20 million. What is the expected net value?
  • A. ₦2 million
  • B. Negative ₦5 million
  • C. ₦7 million
  • D. ₦8 million
B. Negative ₦5 million — (0.50 − 0.15) × ₦20 million = ₦7 million gross; less ₦12 million gives negative ₦5 million. ₦7 million is the tempting answer because it forgets the cost of the response.
2. An AI summary says a paid supplier offer recovers one week, but the client's operating interface, a predecessor of installation, is unavailable until the week after the offered delivery. What should the project director do?
  • A. Fund the offer as summarised
  • B. Ask the model to restate the recovery more confidently
  • C. Check the integrated network, pursue the interface decision and assess any other benefit of earlier equipment separately
  • D. Reject all expediting on the project
C. Check the integrated network, pursue the interface decision and assess any other benefit of earlier equipment separately — In the walkthrough, the integrated calculation and R-03 show no completion improvement once the interface controls installation. Funding on the summary is the tempting error.
3. Your planning team trains a delivery-duration model only on packages that have already arrived and excludes those still in progress. What is the main problem?
  • A. There is no problem if the completed packages are recent
  • B. The model will over-predict durations
  • C. In-progress packages should be recorded with zero duration instead
  • D. The excluded packages are right-censored, so the estimate is biased towards quick packages
D. The excluded packages are right-censored, so the estimate is biased towards quick packages — An in-progress package's duration is not zero and not yet fully observed. Recording zero duration is the tempting fix and makes the bias worse.

Today’s Assignment

Write Section 13 of your plan: the long-lead forecast for one item in your decision chain, the economics of a response and how technical control is kept during recovery.
  • For one long-lead item your company buys, supplies or depends on, list its milestones, the evidence you hold for each and the latest useful intervention. Say honestly whether your company tracks evidence or a repeated promised date.
  • Define the forecast event, decision horizon, issue cut-offs and the transparent comparator any model must beat, or state the simpler service.
  • Test one response with your own assumed figures, labelled assumed: cost, probabilities with and without the response and eligible contractor consequence. Show the arithmetic and one sensitivity case, and keep client production value separate.
  • Name the common causes this item shares with other packages, and show prediction, response, schedule consequence and commitment in separate records with named owners.
  • State what substitution, partial shipment or altered testing would require before a recovery counts as feasible. 400–900 words.
YOUR 90-DAY PLAN

Section 13 — The long-lead forecast, the response economics and technical control in recovery.

How this is marked — the rubric
Tracks the item through evidenced milestones with a latest useful intervention, not a promised date25
Defines event, horizon, cut-offs and a transparent comparator, or justifies a simpler service20
Shows correct expected-value arithmetic on labelled assumed inputs with a sensitivity case and eligible contractor consequence only25
Separates prediction from response, names common causes and preserves technical control in recovery30
Total · pass mark 70100
Submitting session 13
Write your answer in Word or Google Docs. The first line must read Session: 13, or the submission is refused. Upload the Word file (.docx), or switch to the Google Doc tab and paste your document’s share link.
Marked against the rubric above and returned as a written letter, saved to your progress page. Do not submit until you have passed the self-check above.
ASSESSMENT NOW OPEN

Pop Quiz 3 — Constructability, Procurement and Long Lead opens now. Sit it before you start the next session — it is how you find out whether the doctrine landed, and the practice papers may be resat as often as you like.

5 questions · 10 minutes · covers Sessions 11–13 · pass mark 70
Sit Pop Quiz 3 →Closed book, no notes, and Coach Ritchie switches himself off while you sit it.
Capabilities Covered
Evidence-based expeditingForecast designResponse economicsIntegrated network checkCommon-cause riskTechnical control in recovery
Tick it once you have submitted.Next: Session 14 →
Part Three · Procurement, Fabrication and SupplySESSION 14
Fabrication Quality, Traceability, Logistics and Preservation
Session 14 of 30 · 25-Minute Module

Why This Matters

A certificate on file and a weld in the rack are the same thing only if the project can prove the relationship. A truck at the gate and a usable item at the workfront are the same thing only if the conditions between them were met. Chapters 13 and 14 deal with those gaps. AI can read certificates, compare inspection records with the plan and flag overdue preservation. It cannot invent an inspection, a signature or a certainty. This session also works through the book's Case 7, a Bayesian update on defect evidence, so you can ask what a defect statistic actually means.

25-Minute Module
6:00 – 12:00
The denominator, and Case 7
  • A defect rate needs an agreed population, measurement unit and detection process, such as inspected joints or tested length, counted consistently. A higher observed defect count can reflect more inspection rather than worse execution. Compare like work.
  • A Bayesian update, in plain words: start from a stated assumption about a rate, add observations and get a revised estimate that always carries the starting assumption. Case 7 uses assumed inputs:
  • Case 7 step (assumed)Value and recomputation
    Illustrative prior for the defect probabilityBeta(1,1), a modelling assumption, not industry knowledge
    Observation3 defects in 50 independent comparable inspected items
    Posterior under the stated binomial modelBeta(1 + 3, 1 + 47) = Beta(4,48)
    Posterior mean4/52, approximately 7.69%
    Raw observed proportion3/50 = 6%
  • The gap between 7.69% and 6% reflects the chosen prior; report that dependence, not a discovered physical property. Items may not be independently selected, targeted campaigns concentrate on suspect work, and defects can share one cause, such as one material batch. Pooling them as independent can produce overconfidence.
  • THE DOCTRINE

    A low predicted defect likelihood does not establish acceptance of an uninspected item. The model can prioritise competent review; it is not permission to reduce mandatory inspection.

  • Record the data-generation process: which items were eligible, which were inspected, how defects were defined and how repeats were treated.
When someone shows you a posterior to two decimal places, ask for the prior and how the items were chosen.
16:00 – 22:00
Arrival is a sequence of conditions
  • Logistics connects technical release, packing, documentation, transport, clearance where applicable, site access, receipt and preservation. A plan that begins at shipment can miss the upstream conditions that decide whether shipment is possible.
  • Use verified dimensions, weight, route information and handling requirements, not a generative answer in place of a competent transport or lifting assessment. For a Nigerian project, identify the particular port, road, waterway and final access arrangements with their permissions and responsibilities, without assuming every shipment suffers the same delay.
  • Receipt establishes what arrived, where, in what condition and acceptance state. Quarantined, incomplete or unsuitable items are not usable stock. A commissioning delay can create a preservation requirement, and missing preservation instructions are an exception to resolve, not a prompt to invent a generic procedure.
  • For movement scheduling, treat route and equipment suitability as mandatory constraints and optimise cost within the feasible set. With no suitable resource, the model should return an infeasible result rather than treat missing capacity as unlimited. A date chosen without receiving capacity may merely move the queue to the gate.
  • A useful disruption record states cause, affected packages, consequence, response and owner, distinguishing missing documentation, physical damage, unavailable transport and denied access. Keep the original plan beside revised forecasts, or the report can look healthy while construction loses its workfront.
  • THE DOCTRINE

    Tracking a vehicle or container is an information service; linking it to executable work turns it into an operating capability.

Received is a location; usable is a state. Report them separately.
Key Terms

Plausible association

A proposed certificate-to-item match that is likely but not established, distinct from both exact identity and missing evidence.

Prior and posterior

In a Bayesian update, the stated starting assumption about a rate and the revised estimate after combining it with observations. The posterior depends on the prior chosen.

Self-Check — Before You Submit
1. In Case 7, the posterior mean defect probability is approximately 7.69% while the raw observed proportion is 6%. Why do they differ?
  • A. It corrects an arithmetic error in the raw proportion
  • B. The posterior reveals the true physical defect rate
  • C. The difference reflects the chosen Beta(1,1) prior, a modelling assumption
  • D. Three of the fifty items were inspected twice
C. The difference reflects the chosen Beta(1,1) prior, a modelling assumption — Beta(1,1) with 3 defects in 50 gives Beta(4,48), mean 4/52; the raw figure is 3/50. The difference reflects the prior, so treating the posterior as a discovered physical property is the trap.
2. A site manager proposes skipping a required hold-point inspection on a batch of welds because an anomaly model rates them low risk. What should the project director decide?
  • A. Refuse: the model may prioritise review but cannot accept an uninspected item or reduce mandatory inspection
  • B. Approve, provided the score is recorded
  • C. Approve for noncritical lines only
  • D. Approve if the welders have a good record
A. Refuse: the model may prioritise review but cannot accept an uninspected item or reduce mandatory inspection — Chapter 13 and Case 7 say a model can prioritise review but cannot bypass required inspection or accept an individual item. A good welder record is tempting but is not inspection evidence.
3. A transport optimiser returns a plan that assumes a heavy-lift crane is available when no suitable crane can be obtained in the window. What should it have done?
  • A. Chosen the cheapest route and left the crane to site
  • B. Treated the missing capacity as unlimited
  • C. Lowered the handling class to fit available equipment
  • D. Returned an infeasible result for management to resolve through authority
D. Returned an infeasible result for management to resolve through authority — Chapter 14 treats equipment suitability as mandatory; management then changes the plan, obtains resources or revises a commitment. Changing the handling class is tempting but alters an accepted requirement without authority.

Today’s Assignment

Write Section 14 of your plan: how identity survives transformation in your decision chain, how inspection is evidenced as executed, the denominator behind any defect figure and how arrival is tracked to usability.
  • Trace one item in your decision chain from incoming evidence through cutting, fabrication or installation. State where your company distinguishes exact identity, plausible association and missing evidence, and where it honestly does not.
  • Describe how your company shows that an inspection was executed rather than a template filled: points, method, responsible parties, evidence and acceptance basis.
  • Take one defect figure your business reports. State its population, unit and detection process, whether inspection coverage could explain a change in it and, for any model, its assumptions as Case 7 states them.
  • Show how the item is tracked from release to usable at the workfront: pre-shipment conditions, receipt states kept apart from usable stock, the preservation owner and disruptions recorded against the original plan.
  • If your company is a fabricator or vendor, write it for the item you hand over. 400–900 words.
YOUR 90-DAY PLAN

Section 14 — Traceable identity, executed inspection and arrival tracked to usability.

How this is marked — the rubric
Traces identity through transformation and separates exact identity, plausible association and missing evidence25
Evidences inspection as an executed activity with acceptance values left to the project basis20
States a defect figure's denominator and detection process and any model assumptions honestly25
Tracks arrival through conditions to usable stock at the workfront, keeping the original plan30
Total · pass mark 70100
Submitting session 14
Write your answer in Word or Google Docs. The first line must read Session: 14, or the submission is refused. Upload the Word file (.docx), or switch to the Google Doc tab and paste your document’s share link.
Marked against the rubric above and returned as a written letter, saved to your progress page. Do not submit until you have passed the self-check above.
Capabilities Covered
Material traceabilityExecuted inspectionDefect denominatorsBayesian reasoningReceipt to usabilityDisruption records
Tick it once you have submitted.Next: Session 15 →
Part Three · Procurement, Fabrication and SupplySESSION 15
Nigerian Content and Supplier Capability
Session 15 of 30 · 25-Minute Module

Why This Matters

It is tempting to treat Nigerian content as a percentage written into a tender response late in the bid. Chapter 15 treats it as part of project design: capability that must be planned, built, demonstrated and evidenced while the commitment is delivered. The statutory starting point is the Nigerian Oil and Gas Industry Content Development Act 2010, and the book is careful about what it says of it, so I will be too. This session teaches you to plan capability with observable outputs, keep evidence matched to the claim it supports, show supplier-development risk honestly and report durable capability separately from compliance. It closes Week 3, and Practice Exam 3 opens at the end.

25-Minute Module
6:00 – 12:00
A capability plan with observable outputs
  • For each selected scope identify required skills, facilities, equipment, quality controls, supervision and accepted outputs. Assess current capability with evidence. State the development action, owner, cost and timing.
  • Plan element (Chapter 15)The question it answers
    Required skills, facilities and equipmentWhat the scope physically needs
    Quality controls and supervisionHow acceptance is protected while capability grows
    Accepted outputsWhat the team must actually produce
    Current capability, assessed with evidenceWhere the gap is today
    Development action, owner, cost and timingWho closes the gap, at what price and by when
  • For an AI-enabled quality workflow, local capability includes maintaining identifiers, reviewing certificate matches, handling exceptions, administering permissions and reconstructing decisions, the work Session 14 described. Training attendance alone does not establish those abilities.
  • THE DOCTRINE

    Use practical demonstrations. A team should resolve a wrong-revision case, reconcile a material balance and recover an interrupted workflow using realistic records.

  • Record accepted performance and any restrictions.
An attendance list shows who was in the room; a demonstration shows what the room can now do.
17:00 – 21:00
Supplier development, risk and durable capability
  • A supplier-development investment may improve future capability while creating current schedule exposure. Show both. Decide which scope can be developed within the present project and which requires a longer programme or an accepted partnership.
  • THE DOCTRINE

    Do not lower technical acceptance to make the development plan appear successful. A failed acceptance test is information requiring action, not a communications problem to be rewritten.

  • The commercial model should identify who funds development and who owns resulting assets, data and methods under the actual agreements. These arrangements affect continuity after the project ends.
  • Report accepted local outputs, demonstrated competence, succession and supplier performance. Keep statutory compliance evidence separate from the broader management ambition of capability development. Both matter, but they answer different questions.
  • AI becomes part of this operating model when Nigerian teams can use, challenge, maintain and improve the workflow. A permanent dependence on an external specialist for every source update may deliver a demonstration while leaving little organisational capability behind.
If the capability leaves when the specialist's contract ends, it was never yours.
Key Terms

Practical demonstration

A test of capability using realistic records, such as resolving a wrong-revision case, reconciling a material balance or recovering an interrupted workflow, with accepted performance and restrictions recorded.

Durable capability

Accepted local outputs, demonstrated competence, succession and supplier performance that continue after the project, reported separately from statutory compliance evidence.

Self-Check — Before You Submit
1. What does the book state about the Nigerian Oil and Gas Industry Content Development Act 2010?
  • A. That it sets one percentage that applies to every scope of work
  • B. That it sets a statutory framework for Nigerian content, including project planning and specified requirements, and its application is not confined to the operator alone
  • C. That it applies only to operators, not to contractors
  • D. That an AI assistant can determine its requirements for a project
B. That it sets a statutory framework for Nigerian content, including project planning and specified requirements, and its application is not confined to the operator alone — This is the book's statement; requirements for the actual scope come from competent review. The book warns against letting a generic percentage substitute for that review, the tempting first option.
2. At a fictional contractor, a content assistant counts a payment to a supplier's local address as evidence of local engineering. What should the reviewer do?
  • A. Accept it, since the address is in Nigeria
  • B. Accept it if the invoice description mentions engineering
  • C. Remove the supplier from the report
  • D. Treat it as unverified and separate location-of-work, personnel and expenditure evidence before any claim
D. Treat it as unverified and separate location-of-work, personnel and expenditure evidence before any claim — Chapter 15 says a payment to a local address does not by itself establish where engineering was performed or whether capability exists, and an assistant should not convert an invoice description into a verified claim. Removing the supplier is an overreaction; the task is to find the evidence that matches the claim.
3. A fictional supplier in your development programme fails an acceptance test two weeks before a planned report to the client. What is the right response?
  • A. Treat the failure as information requiring action, keep technical acceptance unchanged and show the schedule exposure alongside the development benefit
  • B. Relax the acceptance criterion for this supplier so the plan stays on track
  • C. Reword the report so the failure reads as a learning milestone
  • D. Move the scope to an external specialist permanently
A. Treat the failure as information requiring action, keep technical acceptance unchanged and show the schedule exposure alongside the development benefit — The book says do not lower technical acceptance to make the plan look successful, and a failed test is not a communications problem to be rewritten. Rewording is the tempting option because it protects the report, but it conceals the very information the plan needs.

Today’s Assignment

Write Section 15 of your plan: the capability your decision chain needs, designed into the project, with observable outputs and evidence that matches each claim.
  • Name one scope in your decision chain where your company makes, or should make, a Nigerian content or capability claim. State how its requirements are determined by competent review rather than a generic percentage, and who owns that review by role.
  • Write the capability plan for that scope with every element from today's table, and name one practical demonstration your team would have to pass, using realistic records.
  • List the evidence behind one current or proposed claim (ownership, location of work, personnel, expenditure) and state honestly where your business currently lets one kind of evidence stand in for another.
  • Show the schedule exposure of any supplier development alongside its benefit, who funds it and who owns the resulting assets, data and methods.
  • State how your company reports durable capability separately from statutory compliance evidence. 400–900 words.
YOUR 90-DAY PLAN

Section 15 — Capability by design, observable outputs and evidence that matches the claim.

How this is marked — the rubric
Grounds requirements in competent review of the actual scope, with no generic percentage and no statutory claim beyond the book20
Writes a complete capability plan with an owner, cost, timing and a practical demonstration30
Matches each piece of evidence to the claim it supports and names a current weakness honestly25
Shows development risk and ownership and separates durable capability from compliance reporting25
Total · pass mark 70100
Submitting session 15
Write your answer in Word or Google Docs. The first line must read Session: 15, or the submission is refused. Upload the Word file (.docx), or switch to the Google Doc tab and paste your document’s share link.
Marked against the rubric above and returned as a written letter, saved to your progress page. Do not submit until you have passed the self-check above.
ASSESSMENT NOW OPEN

Practice Exam 3 — Procurement, Fabrication and Supply opens now. Sit it before you start the next session — it is how you find out whether the doctrine landed, and the practice papers may be resat as often as you like.

15 questions · 25 minutes · covers Sessions 11–15 · pass mark 70
Sit Practice Exam 3 →Closed book, no notes, and Coach Ritchie switches himself off while you sit it.
Capabilities Covered
Capability by designObservable outputsEvidence disciplineSupplier development riskDurable capability reporting
Tick it once you have submitted.Next: Session 16 →
← Back to programme home