Planned backlog: do you have a healthy queue, or are you about to run dry?

The planner opens the system on a Monday morning and looks at the queue of fully planned jobs. Some are ready to go the moment a slot opens. Others are planned but still waiting on a shutdown window or a part on order.

Add them up in hours, divide by what the crew can realistically produce in a week, and you have a number that answers a deceptively simple question: at the current pace, how long would it take to work through everything that is already planned?

What it actually measures

Planned backlog is the total hours of planned and ready work divided by the crew's net available capacity per week, expressed in weeks. Planned work has been through formal planning, with labour, materials, tools and safety requirements all identified. Ready work has gone a step further, with materials procured, waiting only for a slot on the schedule. Unplanned work, and work not yet planned, are left out. One point of discipline underlies it all: backlog must be counted in hours of work, never in the number of work orders, because a hundred trivial jobs and a hundred week-long ones are wildly different loads on the same crew.

How to work it out

The capacity figure is where the honesty lives, and where most plants fool themselves. Take a ten-person crew on a 40-hour week with a little authorised overtime: on paper that is well over 400 hours of capacity. But two are on holiday, one is on loan to engineering, everyone loses time to training and meetings, and history says emergencies will eat a fixed slice every single week. Strip all of that out and the real, schedulable number is far lower.

Planned backlog (weeks) = (Planned hours + Ready hours) / Net crew capacity per week

Say that leaves about 235 net hours a week. With 250 hours of planned work not yet ready and 700 hours ready to schedule:

Planned backlog = (250 + 700) / 235 = 4.0 weeks

The trap this avoids is brutal. Measured against the paper capacity of 400-plus hours, the same backlog would look like a comfortable two and a bit weeks; a plant that forgets to subtract its reactive work, holidays and meetings can believe it has a fortnight of work in hand when the true, schedulable figure is several times that. Use gross capacity and the backlog always looks healthier than it is.

What good looks like

Here the honest answer is that the experts disagree, and it is worth knowing the spread rather than chasing a false precision. Widely respected guidance variously puts a healthy total planned backlog at two to three weeks, at an optimum of two to four, or at four to six; ready backlog tends to be quoted lower, somewhere between one and four weeks. The four weeks in the example above sits inside the broad band, though it would look slightly heavy to those who prefer two to three. The lesson is not to fixate on a target number but to watch two things: that you are dividing by net and not gross capacity, and that the trend holds steady. The decision the backlog drives is clearer than any single benchmark anyway: drift much above four weeks and the queue is growing faster than the crew can clear it, a signal to add resources, authorise overtime or bring in contractors; fall below about two and the scheduler is running out of prepared work, and the crew will slide back into reaction to fill the gap.

Where it can mislead you

• Put every component in the same units, normally person-hours, before you divide, and count backlog in hours of work rather than in work-order numbers.
• Use net capacity, not gross. Deduct holidays, training, meetings and a realistic emergency allowance, or the backlog will look far healthier than it truly is.
• It fluctuates legitimately around big overhauls and turnarounds, so compare against a rolling average rather than reacting to a single week.
• A very low planned backlog is as much a warning as a high one. It means the planning pipeline is drying up, and reactive work is about to rush in to fill the vacuum.

Planned backlog is the gauge that keeps a crew supplied with good, ready work. Too little, and the scheduler runs dry and the team slides back into reaction; too much, and the queue is outrunning the people meant to clear it. Held in a sensible band, with the trend watched more closely than the exact number, it is the quiet sign of a planning function that is keeping pace with its plant.