The relief valve sizing runs in Excel. It always has. Process updates a flow case and doesn't clock that the relief size on a different tab moved with it. Saves. Exports a PDF for the submittal. The P&ID still shows the old valve, because the piping lead hasn't picked up the change yet. Both files go out on the same transmittal.

Reviewers compare PDF to PDF in Bluebeam. The calc PDF looks fine: clean tables, the layout everyone's used to. What nobody sees is that cell D14 went from 12,400 to 8,900 lb/hr. The difference is a number in a grid, not a cloud on a drawing.

Why spreadsheets slip through

Review culture is built around drawings. A change on a P&ID gets a revision cloud, and reviewers are trained to hunt for the cloud. A changed cell announces nothing. Same font, same layout, same total across the bottom. The number is different and the page looks identical to last week's.

• A formula cell changes value without changing color, so nothing looks different on the printout.
• There are six tabs; reviewers open the summary, not the inputs tab where the edit actually happened.
• A linked workbook breaks silently the moment someone moves a file in SharePoint.
• The filename says Rev B; the author meant to bump it after the Friday edit and didn't.

Where it actually hurts

That valve size on the datasheet wasn't picked by hand. It came straight off the calc: size the valve for the flow on sheet 3. So when process drops that flow from 12,400 to 8,900 lb/hr, the calc quietly sizes a smaller valve, but the edit was upstream and the datasheet still lists the one sized for the old flow. The calc moved; the thing that depends on it didn't. Nobody notices until procurement, fab, or field fit-up, and by then swapping a valve is expensive and explaining why is worse.

The spreadsheet is where most of these mistakes are born, and it's the one deliverable nobody redlines.

How Kord handles it

• Kord diffs the native workbook cell by cell, so a changed input or formula lights up instead of hiding behind a total that prints the same.
• It diffs every tab, including the inputs sheet reviewers never open, not just the summary at the front.
• The calc gets reviewed in the same session as the P&ID and the datasheet, so the change sits next to the documents that quote it.
• The consistency check does the cross-check for you: the calc now sizes a smaller valve, the datasheet still specs the old one.
• Every approval is pinned to a specific workbook version, so Rev B can't quietly mean two different files.

Treat the calc like a deliverable that gets diffed and versioned the same as any drawing, and the drift turns up in review, which is a conversation, instead of at fabrication, which is a change order.