Control while work is running

Project control:
when you intervene, and with what

A plan rarely survives to the end. Project control is the work that follows: measuring what actually happened, judging whether it costs the delivery date, and deciding which lever to pull. This page covers the four control variables of a running project, the threshold at which a decision is made, and the five interventions available in an industrial project.

Linetrack software for controlling running industrial projects

In short

Project control is the intervening half of project management: it keeps the plan valid during execution by judging deviations and adjusting four control variables – dates, capacity, scope and sequence. A decision is due as soon as a deviation consumes a task's buffer or reaches a downstream task. Linetrack calculates beforehand what each option means for the remaining projects – in use at Stadler Rail and Schnaithmann Maschinenbau, among others.

1512

industrial projects run
in parallel

270+

parallel projects at
the largest customer

59 %

named manual status chasing
as a starting problem

10 WD

median, to the first
self-planned project

*Every figure on this page comes from a full survey of all live Linetrack implementations, as of August 2026. Multiple answers were possible for the starting problems.

Definition

What project control means in a running project

Project control has three parts that follow one another. The actual state of each task is measured: reported progress, booked hours, real dates. The gap to the plan is judged – because not every deviation calls for a decision. An intervention only follows when the judgement says the task will not recover on its own. Drop any one of the three parts and you get either action on suspicion or a report that leads nowhere.

This page deals with the intervening part. How planning and control connect as one loop, and what belongs in a complete plan, is covered under project planning and control; which metrics the measurement produces, and where the actuals come from, is covered by the plan-versus-actual comparison. In machinery and plant engineering this is where a promised delivery date is won or lost: the same design engineers, electrical planners and fitters work across several projects at once, so every intervention in one project moves work in the others.

Control variables

The control variables of a running project

Plenty gets measured, little can actually be moved. A running task has four variables you can turn – and each of them is paid for by another part of the project.

Dates. Start, finish and duration of a task. Extending the duration is the most direct intervention – and the one that eats the buffer of downstream tasks fastest. While reserve remains it costs nothing; once the reserve is gone it costs the delivery date.
Capacity. The hours a department assigns to the task. Adding hours only helps if those hours are free elsewhere – which is why capacity planning across all projects is the precondition for using this lever at all, rather than merely relocating a bottleneck.
Scope. What this project actually builds, ships or hands over. Scope is the only control variable that cannot be settled internally: it touches the commitment made to the customer and therefore leaves the project manager's remit.
Sequence. Which tasks run in which order. Swapping two tasks costs no additional hour, but it requires the dependencies to allow it – otherwise the swap only pushes the problem further back, to a place with less buffer.

What is measured and what can be moved are not the same thing. Percent complete, remaining effort and schedule adherence are measurements: they show the state but cannot be changed directly. Anyone trying to steer percent complete instead of adjusting dates, capacity, scope or sequence is correcting the display rather than the cause.

Decision threshold

The deviation at which you intervene

Not every delay is a control case. A task running two days late with three days of buffer often catches up by itself; intervening there spends capacity on a problem that was none. Intervening too late is paid for with the delivery date. The threshold in between can be stated – and it is not a percentage, but a combination of three criteria.

One intervention over time

An example run across seven reporting dates: a task with three working days of buffer drifts out of plan. The columns show the deviation against the plan in working days.

Deviation inside the corridorBuffer used up or exceededBuffer of the task: 3 working days

W 12: deviation of 0 working days against the plan, buffer 3 working days – task on plan.
W 13: deviation of 1 working days against the plan, buffer 3 working days – first deviation, being watched.
W 14: deviation of 2 working days against the plan, buffer 3 working days – second reporting date pointing the same way, the trend is reliable.
W 15: deviation of 3 working days against the plan, buffer 3 working days – buffer used up.
InterventionW 16: deviation of 5 working days against the plan, buffer 3 working days – downstream task reached, intervention decided.
W 17: deviation of 3 working days against the plan, buffer 3 working days – effect of the intervention.
W 18: deviation of 1 working days against the plan, buffer 3 working days – back inside the corridor.

The intervention falls in week 16, not week 18. The deviation is visible from week 13; it becomes reliable in week 14, once two reporting dates point the same way; it is decided in week 16, because the three working days of buffer are used up and the downstream task is reached. Capacity is added there – two reporting dates later the task is back inside the corridor. The figures are a model run illustrating the rule, not a surveyed metric.

The three criteria apply independently of one another. First, the buffer: as long as the deviation is smaller than the task's reserve, it stays inside the corridor and is watched, not treated. Second, the trend: a single report can be a data-entry error or a shifted day of leave – a deviation only becomes reliable once two consecutive reporting dates point the same way. Third, contact: as soon as a deviation reaches a milestone or a downstream task, the decision is made immediately, regardless of buffer and trend. How much buffer a task holds and whether it sits on the critical path can be read off the bar chart – see the page on the Gantt chart.

For the second criterion to be measurable at all, you need fixed reporting dates. A trend across two data points only exists if both were measured by the same rules and at the same interval. A status collected once on a Monday and next on a Thursday produces a curve made of collection differences rather than project reality – and the threshold then fires either too early or not at all. The control rhythm is therefore not a formality but the condition under which the rule remains applicable.

Interventions

The interventions available to you

Five interventions cover the repertoire. The interesting part is not the list but the third column: each intervention is paid for with a different variable, and which one it is decides whether it is bearable in the situation at hand.

Intervention	When it fits	What it spends
Extend the task	The deviation is small and, for now, confined to the task itself.	Buffer of the downstream tasks. The cheapest intervention while reserve remains – and the most expensive the moment it is gone.
Add or reallocate capacity	The effort was estimated correctly, the hours are simply missing.	Free capacity of a department, which is then missing from another project. Without a cross-project view this intervention is flying blind.
Subcontract a work package	No internal capacity is free for weeks and the scope can be cleanly separated.	Budget and coordination. A subcontracted task needs a handover, an acceptance step and a reporting route, otherwise it is a bar without content in the plan.
Swap the sequence	Another task can be pulled forward and the dependencies allow it.	Sequence stability. Every reshuffle changes the assumptions downstream departments built their own planning on.
Cut scope or defer the project	The other levers are exhausted and several projects compete for the same department.	The commitment to the customer, or another project's date. The only intervention that leaves the project and belongs with management.

All five interventions reach beyond the project itself as soon as departments are shared across projects. That is why Linetrack calculates before the decision rather than after it: the option is simulated, its effect appears as a spike in the capacity histogram of the affected department, and only the chosen option goes back into the plan. What an intervention means for the remaining projects is covered by multi-project management.

Roles and escalation

Who decides the intervention

An intervention is a decision, and decisions sit with roles. The project manager owns milestones, phases and the targets per department. The department head plans within those targets and solves whatever his area allows – he cannot move the targets themselves. Team members report progress on their own tasks and change no dates; where that actual state comes from is covered by task management.

Whatever the department head cannot solve within his area escalates to the project manager – visibly, not as a quiet correction to his own plan. The project manager then sees what the options mean for the other departments, plays the scenarios through, and takes the calculated options into the next meeting. How dates and capacity come together on one data foundation for this is covered under scheduling and capacity planning.

Project management dashboard with Gantt chart and capacity histogram as the basis for control decisions

An honest boundary

What Linetrack does not do

Linetrack does not clear an overload by itself. The system shows that a department is above its capacity, calculates the dependencies through and quantifies what the options cost. Which one is carried out is decided by the project manager. There is no automation that moves tasks on its own – and that is deliberate: an automatic replan nobody ordered gets reversed in the next meeting.

Linetrack is neither an ERP nor a time-tracking system. Actual hours are booked where they are booked anyway and arrive through an interface; master data, purchasing and invoicing stay in the ERP. Which systems are connected, and how the data flows, is covered under integrations.

No claims management, no forensic delay analysis. Linetrack documents when which status was reported and marks the critical path. The retrospective delay and variation analysis required for a contractual claim is not part of it – anyone who needs that needs a separate process for it.

And project control is not project controlling. Controlling measures and puts plan against actual; control decides what happens next. Linetrack supplies the figures for that decision, not the verdict.

FAQ

Frequently asked questions

Still have questions?

Book a slot with our team.

What is project control?

Project control is the intervening half of project management. It has three steps: the actual state of each task is measured, the deviation from the plan is judged, and where the judgement demands it, one of the four control variables is adjusted – dates, capacity, scope or sequence. Without the third step you get a report that leads nowhere.

What are the tasks of project control in a running project?

Four: keep the actual state current, judge deviations against buffer and trend, calculate the options together with their effect on the remaining projects, and write the decision back into the plan. The last step is the one most often skipped – a decision that only exists in the meeting minutes changes neither the plan nor the department's capacity situation.

Which control variables can a project manager change?

Four: the dates of a task, the capacity assigned to it, the scope and the sequence of tasks. Everything else is a measurement. Percent complete, remaining effort or schedule adherence show the state but cannot be adjusted directly. Scope is the only one of the four that cannot be settled internally, because it touches the commitment made to the customer.

At what level of deviation should you intervene in a project?

Three criteria decide, independently of one another. First the buffer: as long as the deviation is smaller than the task's reserve, it is watched. Second the trend: a deviation becomes reliable once two consecutive reporting dates point the same way. Third contact: as soon as a milestone or a downstream task is affected, the decision is made immediately. A fixed percentage is useless as a threshold, because two days of delay are harmless or date-critical depending on the buffer.

What is the difference between project control and project management?

Project management is the whole discipline – from clarifying the order through planning and control to project closure, including roles, communication and risk. Project control is one task within it, the one that happens during execution: it intervenes in a project that is already running. Talking about project control means the decisions taken in operation, not the methodology before them or the evaluation after them.

Who decides an intervention – the system or the project manager?

The project manager. Linetrack does not clear an overload by itself and does not replan automatically. The system shows the bottleneck, calculates the shift through the chain of dependent tasks and quantifies what each option costs. The department head solves what he can within his targets; anything beyond that escalates visibly to the project manager. Linetrack supplies the figures, not the verdict.

What does a system need in order to calculate an intervention before the decision?

Three things. It must know the dependencies, so that a shift runs through the downstream tasks instead of merely stretching one bar. It must hold dates and capacity on the same data foundation, otherwise it shows dates nobody can deliver. And it must calculate across projects, because the capacity an intervention spends is missing from another project. Which methodical approaches sit behind this is set out in the article types of project control compared.

Learn more