How many hours do subcontractors spend on submittals?

You can find figures for this. We are not going to repeat them, because we cannot trace them to a method we would defend, and a statistic with no source behind it is worth less than nothing on a page that is asking you to trust everything else on it.

The honest position is narrower and more useful: we know what one office spent, we know what it spends now, and we know how to find out what yours spends.

Patriot Pipeline assembled construction document packages by hand. That work took hours a day. The tool we built does it in under a minute. Those are the two figures, they are approved, and they are the whole of what we can evidence.

Note what is not claimed. Not a percentage saved across a year, not a headcount avoided, not a figure for submittals specifically. The work was document package assembly, which is the same kind of task and is not the same word, and stretching it would be the easiest overstatement on this site to make.

• Trade. An electrical sub with hundreds of product submittals is running a different process from a grading contractor with a handful.
• Who reviews you. One general contractor with a strict format and a slow reviewer can cost more than the rest of your customers combined.
• How much is rework. Two offices with identical volume can differ by a factor of two, entirely on how often packages come back.

Any single industry-wide hours figure has averaged across all three, which is why it cannot tell you anything actionable about your own week.

Have the people doing the work log it as they do it, not from memory at the end. Separate assembly from chasing, because they are different problems with different fixes. Count the rework separately again, because a resubmittal is not half a submittal, it is a whole one plus the first attempt.

A week is enough to get a defensible number and short enough that people will actually do it. A month of remembered estimates is worse than nothing, because it feels like data.

A single average is the least useful form this can take, because the distribution is what you act on. Two offices logging the same mean can be in completely different trouble: one steady at four hours every day, the other idle for a fortnight and then losing a full week to a single deadline.

So keep the spread. A range with the worst week named, and a note of what made that week worst, tells you where to intervene. A mean tells you nothing except that some work happened.

Watch the outliers rather than discarding them. In most offices the extreme entries are not noise: they are the rework, and the rework is the part worth removing first.

The first week of an install is a baseline, and it exists for exactly this reason: the only number that can hold up a guarantee is one taken from your office before anything changed.

It is also the part we will not shortcut. A guarantee measured against a figure somebody estimated afterwards is not a guarantee, it is a negotiation waiting to happen.