Projects

Why sharing people across projects slows them all down

Mohit Saini · 6 Aug 2018 · 3 min read

Why sharing people across projects slows them all down

In organisations where several projects share the same people (engineering firms, IT teams, R&D centres, construction companies, product development groups), on-time delivery against the original promised date is often shockingly low. The reported performance may look better, but only because it is measured against revised dates, or because scope was quietly cut to hit a deadline.

The usual diagnosis: lack of visibility

Because delays are often discovered close to the due date, management concludes that the problem is visibility and control. So companies invest in project software, build more detailed plans and add more milestones.

Then uncertainty strikes during execution. One project slips. Because people are shared, that slip delays other projects too, and the delays multiply as they compete for the same resources. Before long, many projects are late at once, management attention is spread across too many crises, and issues wait even longer to be resolved.

Many project managers conclude that planning is a waste of time. Halfway through a project, they abandon the plan and run on intuition.

Others try a different tactic: start projects as early as possible, to give them more time. That adds load on shared people, and makes the next problem worse.

Two assumptions behind the chaos

Critical Chain Project Management, developed by Dr. Eliyahu M. Goldratt, traces the mess to two assumptions:

Assumption 1: “Start projects as early as possible, so they finish early.” Starting everything early means many projects are open at once. Shared people end up switching between tasks across projects. Each switch wastes time and capacity, and every project takes longer. Starting earlier makes finishing later.

Picture one engineer with three projects, A, B and C, each needing 10 days of her work.

  • One at a time: A finishes on day 10, B on day 20, C on day 30. The average project finishes on day 20.
  • All started at once, switching in five-day slices: A gets days 1–5, B days 6–10, C days 11–15, then A again, and so on. A finishes on day 20, B on day 25, C on day 30. The average project now finishes on day 25. Nothing finished sooner; two projects finished later.

And that is before counting the cost of switching. Every time she picks a project back up, she has to recall where she left it, re-read her notes, and chase an input she asked for a week ago. Add even one lost day per switch, and every project slips further, with C now finishing well after day 30. Starting all three on day one didn’t give anyone more time. It took time away from everyone.

Real teams have many people, many projects and hand-overs between them, so the effect multiplies. One person’s switching delays the hand-over to the next person, who by then has switched to something else.

Assumption 2: “The only way to finish the project on time is to finish every task on time.” When each task estimate becomes a deadline, people build safety into every task. That safety is wasted through late starts and work expanding to fill the time available. Delays are passed down the chain; gains are not.

What changes with Critical Chain

In planning:

  • Since nobody can foresee exactly what will go wrong, there is little point scheduling every task to a fixed date.
  • The safety is moved from individual tasks to where it protects the project: a project buffer at the end of the longest chain, and feeding buffers where side paths join it.

In execution:

  • The number of projects open at once is controlled, so shared people aren’t spread across too many.
  • The buffers are visible and actively managed. When a project starts eating its buffer faster than it progresses, that is the signal to intervene, early.

Organisations that make this shift typically report on-time delivery rising sharply, and noticeably more projects completed with the same people. Quality tends to improve too, because the pressure to cut corners near the deadline largely disappears.

In a multi-project environment, the fastest way to finish projects is to start fewer of them at a time.

Related reading

Mohit Saini

Strategy, supply chain & sales transformation consultant. About · Book a discussion