Batch Sheets That Survive a Surprise Inspection

The inspector arrived at 09:20 without an appointment, asked for the records behind a specific load that had left eleven days earlier, and gave the site supervisor twenty minutes to produce them. What came out of the office was a clean, typed spreadsheet with every field filled and no gaps. What came out of the mixing shed was a clipboard with coffee rings on it, four sets of initials and a note that said “paddle noisy, told maintenance”. The typed sheet was rejected within five minutes. The clipboard was what saved the file.

That is not an argument against tidy records. It is an argument about when they were written.

What an Inspector Is Actually Testing

The question underneath every request is not “do you have a record” but “could this record have been produced at the time, by the person named, describing the batch that actually went out”. A spreadsheet re-typed at the end of the week fails that test on its face: it is a summary, not a record, and there is nothing to tie it to the physical event.

What ties a record to an event is small and unglamorous. A time written in the same pen as the rest of the page. A polymer lot number copied off the bag at the moment of use. A signature with a printed name next to it. A retained sample with the same batch number written on the lid. A temperature that was actually read rather than estimated.

The Fields That Earn Their Place

One page per batch, no exceptions, and no batch closes without a page.

• Date, start time, finish time, shift.
• Source: which sump or tank, and the volume drawn.
• Polymer: grade, lot number, bags or kilograms issued.
• Dose: kilograms per cubic metre, and the calculated total, written before dosing rather than after.
• Mixing: minutes, and any interruption with a reason.
• Liquor temperature and pH at the start.
• Test result: which test, the reading, the time, and who read it.
• Operator initials and printed name. Supervisor check on the same page.
• Retained sample number, matching the label on the jar.
• Free text: anything unusual. This is the field that saves you.

The free text field looks like an invitation to write nothing. It is the one that carries the most weight, because it shows the record was written by someone watching the batch rather than reconstructed afterwards.

The Failure That Gets Found

A site produced a beautiful digital record covering fourteen months with a 100 percent completion rate. During a review, a batch was cross-checked against the weighbridge log and the tanker’s own timesheet: the load left at 13:10, but the batch record showed mixing completed at 15:40. Two operators had signed for a shift on which the day book and the gate log both placed them elsewhere.

The cause was ordinary. When the digital system was introduced, the crew back-filled three weeks of history to make the record look complete from day one, and from then on they filled it in at the end of each shift, from memory, at speed. Every number was plausible and none of it was contemporaneous.

The consequence was worse than the original gap would have been. A three-week gap in a record is an administrative failing. A record that can be shown to be reconstructed is a credibility problem that spreads to every other record the site holds.

Making the Paperwork Fast Enough to Do Properly

If the form takes four minutes, it will be filled in at the time. If it takes twelve, it will be filled in later. The practical target is one side of A4, pre-printed with the fields, with the polymer lot and the sump number pre-filled at the start of the shift and only the batch-specific fields left blank.

Attach the record to the physical batch: a batch number painted or taped on the tank, matching the number on the sheet and on the sample jar. When the inspector asks about a load, the trail from weighbridge ticket to batch number to sheet to sample should take under two minutes to follow. If it takes longer than that, the file is not really organised, it is just complete.

Retention and the Sample

Keep the retained sample for the period your acceptance arrangement specifies, in a labelled jar, out of sunlight, and log its disposal. The jar is worth more than the sheet in a dispute, because it can be re-tested by someone else. A sheet with no sample is a claim. A sheet with a sample is evidence.

The record-keeping section of the operating guide has a one-page template that covers these fields without turning the shift into a clerical exercise.

One last habit worth building: at the end of each week, pull three random batch sheets and try to reconstruct the load from the surrounding paperwork without looking at the sheet. Every one you can reconstruct is a record. Every one you cannot is a form.