Recovery plans can consume the remaining margin

Recovery should describe an executable intervention rather than an instruction to regain time. Shortening a task in the model is an analytical scenario. Delivering that shorter duration requires a mechanism: changed resources, improved release completeness, a different sequence, reduced scope or another method. The recovery record should state the mechanism and evidence before using the revised date as a promise. Otherwise the schedule becomes a target generator detached from the work required to achieve it.

Crashing increases effort or capacity on an activity to reduce elapsed duration. Its benefit is not automatically proportional to added resources. Some work cannot be divided, and additional teams can create supervision, access or quality constraints. A supplier’s acceleration offer should therefore explain the changed production basis, available capacity and acceptance implications. The Delivery atlas can test a revised duration assumption after that basis is defined. It cannot establish the productivity gain or calculate the intervention’s cost.

Concurrency changes sequence or release boundaries. It can be useful when a successor can proceed safely from partial information. It can also expose work to redesign, rework or incompatible interfaces. The dependency register should identify which condition is relaxed, what scope is permitted and who accepts the exposure. A link should not be removed merely because its deletion improves the finish. The recovery method needs a causal basis that the recipient can rely upon.

Scope reduction can improve delivery when the revised outcome still meets the decision objective and admissibility requirements. It should be recorded as a change in obligation, not disguised as faster performance against unchanged scope. A deferred optional feature may be a reasonable response; an omitted mandatory test is not made acceptable by a favourable date. The authority should distinguish negotiable scope from noncompensable conditions before comparing recovery options.

Recovery can move the critical path. The duration-switch study improves the supply assumptions and exposes receiving works as binding. Its conditional completion is 17.00 weeks. The result limits the schedule benefit of further supply-only action under the same network. An integrated recovery plan should therefore recompute after each material change and review competing branches. Local acceleration can be real without delivering an equal project saving.

Concurrent work can create a contingent rework obligation. If assembly begins before all design information is fixed, a later change can affect completed work. The recovery decision should identify the possible rework scope, detection point and stopping condition. The app’s acyclic network does not represent rework loops or stochastic branch outcomes. A simplified scenario can add an explicit rework task, but it must be labelled as a separate assumed case. It does not estimate the probability or full consequence of rework.

A recovery action can create new interfaces. Splitting a supplier package may allow earlier partial delivery but require configuration control and additional receiving checks. Adding another contractor may reduce resource waiting while increasing coordination needs. Moving tests earlier may require different evidence or accepting authority. These obligations should be included in the revised perimeter. A plan that counts only the time saved and omits the integration work is systematically incomplete even if each local estimate is credible.

Recovery should preserve quality and safety conditions. They are not generic preferences to be traded against a schedule gain. The actual applicable requirements determine the admissible set of methods. The control record should identify approvals or competence needed before the intervention begins. A sponsor’s desire to accelerate does not supply technical permission. Conversely, unnecessary administrative constraints should be challenged through the appropriate authority rather than treated as immutable physical laws.

A proposed action needs a readiness date of its own. Additional resources may require mobilisation, qualification or procurement. A changed method may need design review or equipment modification. If these enabling tasks are omitted, the recovery scenario can overstate the available saving. The complete network should include material enabling work or explicitly state that it is outside the study. A late decision can make an otherwise credible intervention no longer executable within the remaining window.

Recovery monitoring should test the mechanism rather than only the revised finish. If the action depends on complete design release, monitor release completeness. If it depends on supplier capacity, monitor the confirmed production arrangement. If it depends on a partial interface, monitor its condition and residual obligation. These leading checks reveal failure earlier than waiting for the final milestone to slip. They also let the authority stop an intervention whose exposure has exceeded its approved boundary.

A fictional assembly contractor proposes to begin before the final interface drawing is released. It can perform some work from stable information, but another portion depends on a pending dimension. The original finish-to-start relation is too coarse to distinguish those scopes. Deleting it entirely would overstate readiness; retaining it unquestioned could miss a valid opportunity. The recovery review defines the unaffected work and the uncertain boundary before changing the network.

The technical authority determines which work can proceed under actual requirements. The authorised partial scope has an explicit stop point before irreversible interface-dependent work. The contractor states resources and acceptance evidence for that scope. The design owner states when the pending information can be accepted. This creates a causal basis for a partial-release scenario. The paper does not invent its durations or probability of rework. Those would need a separately frozen numerical study and evidence if relied upon.

The manager identifies contingent rework. If the pending dimension changes beyond the permitted boundary, completed work may need modification. The decision should state who monitors the change, what triggers a stop and which exposure is accepted. A generic contingency added to assembly does not explain this branch. The acyclic app can represent a declared correction scenario, but it cannot simulate a loop or infer its likelihood. The scenario's scope remains narrower than the real adaptive process.

The sponsor compares the opportunity with a wait-for-release alternative. Earlier activity may preserve resources or shorten completion, but it can also increase sunk exposure. The comparison includes enabling work, inspection and possible interface correction. The critical-path result helps determine whether the permitted scope affects accepted completion. It does not decide whether the exposure is admissible or worth accepting. Those judgments remain with the actual authorities and evidence.

The partial release is communicated to the recipient in exact terms. A meeting summary saying assembly approved would be too broad. It should state which area, configuration and activities are permitted and where the hold remains. The programme then distinguishes partial work from full assembly acceptance. This prevents activity from being reported as a completed predecessor while the final interface remains unresolved. Clear scope can make a recovery action both faster and more controlled.

The post-action review retains the release basis and the eventual interface disposition. If no rework occurs, that favourable outcome does not prove that unrestricted overlap would have been safe. If rework occurs within the accepted exposure, the decision may still have been reasonable. The record supports learning from the mechanism rather than from hindsight alone. The transferable lesson is to divide scope at a credible release boundary instead of treating concurrency as a binary permission to remove logic.

In a fictional recovery workshop, the sponsor asks for an earlier accepted installation. The supplier offers acceleration, the works team offers an overlapping receiving release, and the operator warns that test preparation remains incomplete. The manager treats these as separate proposed mechanisms. Combining them immediately into one optimistic network would make it hard to identify which assumptions drive the result or which intervention must be secured. The first analysis preserves the reference and evaluates declared changes.

The supply-only study shows path switching. The team then asks whether receiving can credibly improve, and whether earlier assembly would be useful given the operator’s test readiness. This second question is outside the frozen supply comparison. It requires a defined test-preparation obligation, resource basis and acceptance condition. The manager should not extrapolate a project saving from a chain of local promises. The integrated plan needs the complete release path to the accepted outcome.

The recovery workshop completes an irreversibility comparison. Additional inspection preparation can be stopped with the unused work preserved. A production-location change commits qualification and logistics work before the first accelerated unit is accepted. A partial receiving release commits only the explicitly stable foundation work until the pending interface is resolved. These alternatives consume reversibility at different points even if their proposed finish dates are similar. The sponsor asks which evidence must precede each commitment and which later finding would be too late to reverse it economically or physically. The record does not price those exposures; it identifies the points at which a recovery decision stops being a reversible planning choice.

The fictional recovery instruction records METHOD-PARTIAL as its selected boundary: preparation of the stable receiving area is permitted; interface-dependent fastening remains held; the pending drawing revision governs the next release. The contractor acknowledges the stop point in its work instruction, while the design owner links the unresolved dimension to the receiving record. This is more specific than authorising acceleration in general. If the drawing changes outside the stable area, unaffected preparation can continue. If the change reaches that area, the instruction itself identifies the work to suspend and inspect. The monitoring object is the interface footprint, which lets the team respond before a final-date variance reveals the exposure.

The post-action instrument compares the chosen mechanism with its closest rejected alternative. It records whether the stable work remained usable, whether the pending interface entered the permitted footprint, whether extra inspection was needed and whether the governing downstream branch changed. A favourable finish with unusable early work indicates that another mechanism produced the outcome. An unfavourable finish with preserved useful work may show that the intervention helped while a separate constraint dominated. The review therefore avoids scoring the method from the final date alone. It preserves a causal fingerprint that the next recovery workshop can compare with its own work boundary, rather than treating successful acceleration as a transferable productivity rate.