A repaired part carries three histories and they are almost never the same. The part has its own: serial number, time since new, time since overhaul, the release tag it shipped out on. The aircraft has a second one: which tail, which position, why the part came off. The shop has a third: what the bench actually found, what got replaced, who signed. Aerospace MRO traceability fails between those three records, because nothing connects them except a person who remembers. I spent an afternoon in a records office at an accessory shop last spring while a clerk worked one hydraulic actuator across three windows. The operator’s maintenance system had the unit at eleven hundred hours. The shop’s work order had it at seven hundred. The release tag in the folder said nothing about hours at all, because it does not have to. She had all three documents open and no way to know which was wrong. So she called a man who had been at that shop nineteen years, and he told her from memory that the unit had sat in a stores cage for most of a year while somebody on the operator’s side kept counting against the tail. Two minutes on the phone. Problem solved. Nothing written down. The three records a repaired part has Each of the three records is complete and correct. They are just about different things. The part’s history is about an object. Part number, serial number, accumulated hours and cycles, life limits, the airworthiness release it travels on, an 8130-3 or an EASA Form 1 depending on where you are. This record travels with the part. It is the one an auditor asks for first. The aircraft’s history is about a position. Number two engine. Left main gear. The logbook is not organized by serial number. It records whether the position was serviceable, whether the directives applicable to it were complied with, and why the part was removed. Removal reason is the single most valuable line in the whole file and the first one to get reduced to a dropdown code. The shop’s history is about a job. A work order, a router, the steps in order, the inspection findings, the non-destructive test results, the sub-components consumed, the deviations, the scrap, the sign-offs. It holds the most detail of the three and shares the least. Almost none of it leaves the building. What leaves is a tag. Object, position, job. Why aerospace MRO traceability breaks between them Aerospace MRO traceability is the ability to show, from records alone, what a part is, where it has been, what was done to it, and by whom. No shared identifier. The part record is keyed on part number plus serial. The aircraft record is keyed on position. The shop record is keyed on a work order. There is no shared identifier, so a person matches the records by reading two screens. When a part comes back modified with a new dash number, even matching the part to its own earlier record breaks. Time is counted differently. Hours accrue against the aircraft. Calendar time accrues against the part. Shelf time counts toward some limits and not others. Zeroing after an overhaul is a decision somebody makes and records in one system, and the other two are never updated. Different levels of detail. The operator sees a line-replaceable unit. The shop works at sub-component level. When a bearing, a seal carrier or a circuit card gets swapped at the bench, the aircraft record has no field for it, so once the box is closed that change is recorded nowhere outside the shop. The three records update at different times. The tag ships with the part. The work order closes weeks later, after billing. The operator updates the aircraft record when the unit is installed, which might be six months out. Three records, three moments, and no one moment where all three are true at once. Copies. Every time a record is copied into another system, the copy starts to differ from the original. A PDF in a folder, a line retyped into a spreadsheet, a number read aloud on a call. A reconciliation test you can run this week Pick one part that came back through a shop in the last quarter. Give yourself thirty minutes. Records only, no phone calls, no walking down the hall. Answer these six: Why did it come off? In the words the person who removed it used, not the code it was mapped to. Then find the shop’s findings text. Do they describe the same failure? What was it carrying when it came off? Hours and cycles, and which of your systems is the authoritative one for that number. If two systems disagree, note the gap. What changed inside it? List the sub-components replaced at the bench. Then check whether any record outside the shop shows that they changed. What went out versus what came in? If the part number changed, find where that change is recorded on the aircraft side. Who inspected it, and what did they find that is not on the tag? There is almost always a finding that mattered and did not make the release document. Can you get from the tag back to the previous shop visit without opening an email or a shared drive? Now score it. Every question that would have required a phone call is a divergence. Write down the name of the person you would have called. That list is who your traceability actually depends on, and it is shorter than you would like. I have watched quality managers run this on a Tuesday morning and go quiet. The records exist. The connections between them do not. How to keep one traceability record for a repaired part Most of what looks like an aerospace MRO traceability problem here is really a question of which record counts as the part. Settle that first. Keep one record per physical part. The serial number identifies the object. A part-number change, a modification, an overhaul, a life reset: all of those are events recorded against that one record, not new records that replace it. That is what back-to-birth traceability depends on. The moment a modified part becomes a new record, you have started a second history. Record events, not states. “Hours since overhaul: 0” erases the reason. “Overhauled on this date, under this authority, hours reset per this instruction” keeps it. States overwrite. Events accumulate, and accumulation is what an investigation three years from now actually needs. Keep the findings where the measurements are. Inspection findings belong on the router step where the inspector wrote them, attached to the measurement, not summarised into a paragraph in a separate document that nobody can trace back. Hand off links, not copies. When a record has to go to the operator or to the next shop, it should point back to the original rather than carry a snapshot of it. A copy is a second record that can change on its own. Protect the removal reason. It is the only line that connects the aircraft’s story to the part’s story. Let people write it in their own words and keep the dropdown as a secondary field, not the other way around. What a record must contain and how long you keep it is governed by the applicable rule text and your quality organization. None of this changes that. It is about whether the three records you already keep can be read as one. The man who reconciled that actuator is going to retire. When he does, the three records will still be sitting there, each one accurate, each one about something slightly different. The thing that held them together will be gone, and nobody will be able to point at what was lost. That is the part worth writing down while he is still in the building.