Four reports, forty per cent of the hours

Counting the thing nobody had counted

The brief was literal: list every recurring report created at the plants. Eleven sites across eight countries, banded small, medium and large, each surveyed and then interviewed, with every report captured against the people involved, the hours per instance, the frequency and who receives it.

Annual hours were then calculated as hours multiplied by people multiplied by frequency, with the frequency assumption printed on every plant page so the arithmetic can be checked. All of it is self reported by the sites, which is stated here for the same reason it was stated there.

The total came to just under 17,000 hours a year across the eleven, which is roughly nine full time people doing nothing but writing reports.

Size does not predict reporting load, and the gap is enormous

The intuition is that a bigger plant writes more reports. The data says the opposite. The heaviest site is a small one at 5,956 hours a year. The lightest is a large one at 128 hours across four reports. That is a factor of 46.

Within the small band alone, one site spends almost twenty times what another small site spends. Same company, same kind of plant, twenty times the reporting effort.

The concentration inside each plant is just as sharp. At one site the top four reports account for more than 85 per cent of the total. At another, a single daily report accounts for more than 80 per cent. Reporting effort is not spread thin, it is piled up in a handful of documents.

Where the time actually goes, and where it does not

By report type, inbound and outbound logistics is the heaviest at 6,764 hours across twelve reports, an average of 564 hours each. Inventory is second at 2,586 across ten.

Environmental, health, safety and quality reporting is the most numerous category by a long way, at forty reports, and consumes 925 hours, an average of 23 hours each. Across the small plants there are 53 such reports and they take 1,300 hours between them.

That inverts the usual assumption. The reports people complain about are the compliance ones, because there are so many of them. The reports that cost money are a dozen logistics documents that nobody mentions.

Paper first, and one site that already solved it

At two sites, operational data is written on paper and then transcribed by hand into multiple systems. At one of them the sequence is paper, then a spreadsheet, then manual entry into the enterprise system, on a site described as having almost no infrastructure for getting data out of the process automatically.

One site is already the internal benchmark. Its tank level report, the direct equivalent of a daily inventory report that takes 500 hours a year elsewhere, is fully automated and takes zero. Its site manager put it plainly: everything that works here should also work at the less complex sites.

The named root cause is not effort, it is tool use. The enterprise system can retrieve the data, but in several cases people did not know how to use it, and it does not return data in a readable format, so it gets extracted into a spreadsheet before anyone can act on it. One plant manager offered the rule the whole study points at: when a new tool arrives, two old ones should leave.

What was identified, and what it is not

Roughly 6,200 hours a year were identified as removable, on the stated assumption that automated reports take zero hours and improved reports take half their current time. Those two assumptions do a lot of work and they are on the page for that reason.

Redundancy turned out not to be the story. Only one genuinely redundant report was found without a plan to remove it. The other six found were already being removed, and three plants reported none at all.

Three recommendations followed. Use the enterprise system to its potential, which mostly means training rather than software. Attack logistics and inventory reporting first, because that is where the hours are. And reduce the number of tools, because the current process at most sites is extracting from several systems, reconciling the formats and checking for consistency before any report can be written.

Nothing here was implemented during the sprint, and no hours were saved. 6,200 is an identified figure, not a banked one, and it is the honest limit of what a ten week count can produce.