COA batch drift: 7 suppliers, 14 weeks, purity swing?

Same batch code across March-April is the tell. I track 5 sites, and 3 only update COA when they refresh the whole page. Not proof of anything, but it smells like copy-paste.

Copy-paste is one thing. A 2.6% purity swing under same code is worse. Did you check the lab method date? Some COAs don't list it.

That verify link idea is decent, but half the time the link just opens the vendor's own PDF. Not independent.

I've kept a spreadsheet of public COA dates for 11 vendors since last fall. The vendors with a real third-party portal update within 7-10 days of a new batch label. The ones with PDFs only tend to lag 6-8 weeks, if ever.

That doesn't prove reliability either way, but it separates admins from actual labs. The 96.8 to 99.4 thing could be a method switch, not a bad batch. Ask for the chromatogram and method column. If they won't post that, the COA is mostly decoration. I'm not saying it's fraud, just that public data has limits.

Method switch was my first thought too. But same batch code plus no method date? Thats not a switch, thats lazy.

I've seen vendors post COAs from a lab that dissolved two years ago. Dates on the PDF were edited metadata. Always check file properties.

I’ve been doing the boring version of this—screenshotting the batch sticker next to a $12 kitchen scale and a calendar, not because I think purity is fake but because I wanted to see if my own week-to-week water weight tracked with batch changes. Fourteen weeks in, my weight swung 3.8 lbs between week 3 and week 11 on the same batch code while my waist stayed flat, and the one time purity dipped from 97.2% to 96.1% I actually felt fine, while the “99.4%” batch landed during a terrible

Three batch codes, eleven weeks, one spreadsheet column for “how I actually felt” — that’s the only pattern I trust now. I got weirdly obsessed with COA purity last spring and even color-coded it, but the numbers that mattered were my Sunday waist measurement at the navel and sleep hours: waist dropped 1.25 inches while my scale bounced in a 3.8 lb range, and sleep went from 6 to 7.5 hours.

When I lined that up against the batch numbers, the purity “drift” had zero relationship to how I felt; the only real red flag was a vendor swapping the testing lab name while keeping the same batch code, which looks like a copy-paste refresh rather than a new test. So I’d log the lab name and COA date next to the batch code and otherwise ignore the purity swing unless the method changed too.

I’m the weirdo who logs Sunday morning weight and Friday dinner out, and after 9 weeks my “purity” swings were nothing compared to my water weight—3.2 lbs up after Thai takeout, 1.4 down after a boring chicken-and-veg weekend. That’s why the 96.8 to 99.4 thing on one batch code reads more like a re-uploaded sample or sloppy site update to me than a rel red flag; if the product were actually drifting that much, I’d expect my own week-to-week averages to look way noisier than they do.

The turn for me was realizing I was refreshing pages and stressing about batch codes while sleeping 5.5 hours and barely walking, and when I fixed those two things my waist finally moved half an inch in 3 weeks. Now I just do a monthly check: same day, same time, fasted, and I compare that to how my jeans fit, not to a COA timestamp.

If you want a useful column, try adding sleep and Saturday sodium next to the batch stuff—my data got way less scary once I saw the actual pattern.

I admire your discipline with this. One thing I have added to my own weekly log is the PDF metadata—creation or modification date, authoring lab, and whether the file was replaced at the same URL—because posted timestamps can be edited later.

A re-used batch code with a purity jump from 96.8% to 99.4% feels less like site upkeep and more like a documentation gap unless there is a clear retest or amended COA note. Do you also save the original

I’m less worried about the purity number moving than about the dates. When I see a COA, I now jot down three things: batch code, lab test date, and the date it was posted.

If the batch code is old but the test date is new, that’s probably a retest or lazy upload. If both are old and purity changes under the same code, that smells like copy-paste or drift.

Have you been able to tell whether the 96.8→99.4 shift came with a new test date? That’s the clue I’d chase before assuming the worst.

I’d log the lab and assay method before the purity number. Same batch code going 96.8 to 99.4 could be HPLC vs UPLC, different columns, or a different testing lab entirely.

I started adding COA test date, lab name, and whether it’s a submitted sample vs retained sample. Also snap a pic of the vial cap/lot on arrival—site batch codes don’t always match the box.

If those don’t line up, that’s my red flag, not the swing itself. How are you linking the public COA to the actual vial you received?

Recycled batch codes always give me the ick—fresh date over shiny purity number, every time.

i started hashing the PDF and checking its metadata (producer, created/modified). same batch code with a fresh purity number but a modified timestamp after the batch date screams “regenerated doc” to me.

also: does the COA have a QR/verify link, and does it actually resolve to that lot or just the homepage? that’s been my bigger tell. curious if your 7-supplier sheet has a “verify link works y/n” column yet?

I screenshot the date first—shiny purity means nothing if the batch code's a rerun. 😬

My empty-nester brain now treats COAs like my sourdough starter log—date seen, URL, and a Wayback capture. The thing I’d add: check whether the QR code on the COA resolves to the same lab report each time.

I’ve seen one batch code’s QR lead to a generic lab homepage later, which felt less like site maintenance and more like a quiet swap. Also, does anyone cross-check the vendor’s “posted” date against the lab’s own report date? That gap has told me more than the purity

As an empty-nester who now does a morning walk and a “receipt ritual” with coffee, I log one extra column: testing lab name + method + whether the COA is signed. Same batch code flipping purity 96.8 to 99.4 would bug me less than the lab changing silently. A rerun at the same lab? Probably sloppy

One column I added after getting burned: lab name + assay method, not just purity %. A 96.8→99.4 jump under the same batch code could be a retest on a different column/method or a

My airport layover COA scroll now rivals my sourdough log—rerun batch codes get the side-eye 😅