A purchase record exists to answer one question later: what was this material, where did it come from, and what happened to it. Six records answer that between them, and one of the six is load-bearing for all the others.
The load-bearing record is the lot number
Everything else hangs off it. A certificate without a lot number cannot be tied to a vial. A usage log without a lot number cannot be tied to a certificate. An invoice without one records a transaction rather than a material.
If only one thing is captured consistently, capture the lot. Lot traceability and recordkeeping covers the mechanics.
| Record | What it answers | Keep for |
|---|---|---|
| Purchase order and invoice | Who sold it, when, at what price | As long as financial records |
| Lot-matched certificate | What the material was when tested | Life of the material, plus your record period |
| Receipt and inspection note | What condition it arrived in | Life of the material |
| Storage log | What conditions it was held at | Life of the material |
| Usage log | What was drawn, when, by whom, for what | Your record period |
| Disposal record | That it left inventory, and how | Your record period |
Six records. The certificate and the two logs are the ones most often missing.
What a receipt record should capture
- Date and time of arrival, because a temperature question is always a question about elapsed time.
- Condition of the packaging and whether any cold chain was intact, which cold chain covers.
- Lot numbers received, checked against the certificates supplied.
- Who received it, so the record has a person attached to it rather than a department.
- Any discrepancy, raised in writing at the time rather than recalled later.
The usage log is the one that gets skipped
It is also the one that makes the rest useful. A log recording that a given lot was drawn on a given date, by a named person, for a named purpose, is what converts a shelf of vials into a documented programme. Without it, the certificate proves what arrived and nothing establishes what happened next.
It does not need to be elaborate. Date, lot, quantity, person, purpose. Five columns.
Retention
There is no single answer and the honest guidance is to pick a period, write it down, and apply it. Financial records typically have their own statutory period. Material records are usefully kept for the life of the material plus whatever period your own programme or any applicable requirement sets, whichever is longer. The decision matters less than it being a decision rather than a habit.
Why your own copy of the certificate matters
A supplier archive is outside your control. Documents move, get superseded, or disappear when a lot is retired, and a link in your records that resolves to nothing is worse than no link because it looks like a record until someone follows it.
Download the certificate for each lot on receipt and store it with the purchase record. It takes seconds and it is the difference between holding evidence and holding a reference to someone else evidence.
Who should be able to reconstruct the record
Someone who was not there. That is the test worth applying, because it is the situation in which records are actually read. A record that makes sense only to the person who wrote it has not been written down in any useful sense.
In practice that means spelling out lot numbers rather than abbreviating them, naming people rather than initials, and recording the purpose of a withdrawal in terms a colleague would recognise a year later.
The failure mode worth designing against
Not missing records. Unlinkable ones. A practice can hold invoices, certificates and a usage log and still be unable to answer the original question, because no single field runs through all three. The lot number is that field, and the cheapest control available is refusing to file any of the three without it.
That rule belongs in an SOP for receipt and storage, where it is enforced at the point material enters.
This sits inside the wider picture of what gets tested and why, which why peptide testing matters sets out across the whole analytical panel.