Projects
Why construction projects run late, and how to win back the time

Large construction and plant-erection projects have a poor track record. Government audits in many countries regularly report delays and cost overruns, and private-sector projects are rarely better. Some organisations have never delivered a large capital project on its originally committed date, in full scope and within budget. Those that claim to be on time are often measuring against revised dates.
Ask project managers why, and they will list the usual suspects:
- delayed decisions and approvals,
- scope changes and rework,
- late appointment of contractors and vendors,
- bad weather,
- too few resources, or poor contractors.
Most of these feel outside management’s control. So is there anything to be done?
“We should have planned better”
The project management body of knowledge offers a vast set of tools and processes, most of them about planning. The implied promise: the better we foresee the future, the better the project will go.
Reality disagrees. However good the plan, uncertainty spoils it quickly. In many organisations, plans are made enthusiastically at the start, go haywire within weeks, and the rest of the project is run on the project manager’s sense of what is urgent that day. Some stop making real plans at all, except to satisfy outside agencies.
If poor planning were really the cause of delays, one painful project would teach everyone to plan better, and performance would improve over time. It hasn’t, not much. So let us look for a different cause.
Where the time actually goes
Take any completed project and split its duration into three parts:
- Actual work on the longest path.
- Waiting and interruptions.
- Rework.
The actual work is typically only a quarter or a third of the total. The rest is waiting and rework. Some of that is genuinely uncontrollable, but a large part is waste that can be controlled, and that time could be used to absorb the real surprises.
Picture a 30-month plant-erection project. Looking back at the critical tasks, the actual working time adds up to about nine months. Another seven months were spent waiting: for approvals, for drawings, for a contractor to mobilise, for materials. Four months went into rework after early starts on unfinished drawings. And the rest was lost to crews and engineers being shuffled between this site and others. If even half of the waiting and rework had been avoided, the project would have finished about six months earlier, with time to spare for the genuinely unexpected.
So why is so much time wasted?
How time leaks out of a project
Estimates become commitments. In an uncertain world, people protect themselves with generous estimates. Yet tasks still rarely finish early, because:
- Work expands to fill the time (Parkinson’s law). This is especially visible before construction starts, in the office phase, when little money is committed and time isn’t watched. Contractor selection, drawing approvals and vendor evaluations drag on. Decisions wait for committee meetings that happen once a month. Civil drawings go through round after round as each stakeholder adds an idea. The actual working time in this phase may be a small fraction of its duration, and yet speeding things up here is cheap compared with later. On site, contractors mobilise slowly and work to agreed norms even when they could go faster.
- Work starts late (student syndrome). Knowing the deadline, people start close to it. Vendor follow-ups intensify only near delivery dates, too late to fix the problems they uncover.
Delays add up; gains don’t. Because of both behaviours, tasks finish on their date at best, or late when hit by a surprise. At every point where paths join, the latest path decides when work can continue. Gains on other paths are lost. Delays accumulate towards the end, and the teams there work under enormous pressure.
Fast-tracking creates rework. When the pre-construction phase runs late, site work is started before drawings and specifications are final, to “save time”. Changes then cause interruptions or demolition and rework on site.
Too many work fronts. The most damaging leak is bad multitasking: too many projects or work fronts open with too few people. Designers switch between projects. Contractor crews are shuffled between sites. Managers’ attention is spread so thin that issues wait days to be resolved. Every work front takes longer than it would on its own.
Four steps to win back the time
Critical Chain Project Management offers a way to stop the waste and keep enough protection for real uncertainty.
1. Limit open work fronts. Release projects and work fronts in a staggered way, keeping a steady number in progress, sized to the organisation’s scarcest resource, which is often management attention, sometimes cash. It feels like slowing growth. In practice, fewer open fronts, fully resourced, finish faster, and the organisation completes more projects in a year.
2. Plan just enough. With fewer fronts open, a plan can actually hold. Build a plan good enough for execution, neither absent nor absurdly detailed. Cut each task estimate to a focused duration, openly accepting that many will be missed. Move the safety into a project buffer at the end of the longest path and feeding buffers where side paths join it. Pooled safety needs less time than the sum of individual safeties. The buffers are meant to be used.
3. Manage execution daily. Since task estimates aren’t commitments, don’t judge people on them. Instead, review open tasks every day: how much time remains, what is in the way, and when the next task can start. Use that to prepare upcoming work and to spot waste immediately. Set priorities by buffer status, so people and management attention stay on the current longest path.
4. Control projects by buffers, not volumes. Construction progress is usually measured in volumes (metres laid, tonnes erected, area completed) or cost booked against plan. These measures can be satisfied by working on non-critical areas while the longest path stands still. Instead, compare progress on the longest path with buffer consumed. Green: progress outpacing buffer use. Yellow: about even. Red: buffer burning faster than progress, so plan a recovery. Review meetings shift from explaining missed milestones to recovering red projects.
Construction projects rarely run out of time. They lose it early, in small pieces, and then pay heavily to buy it back at the end.
Organisations that apply these steps typically finish projects significantly faster, often on the original date, without budget overruns or cut scope.