I’ve been logging public supplier COA pages and email auto-replies for 12 weeks—37 suppliers, 214 data points. Median first human reply is 5 days; 6 vendors never replied.
HPLC purity on posted COAs ranges 92.1–99.4%, and three show mass spec values that don’t match the parent ion. Is this just poor documentation, or a reliable red flag? I’m not naming names, just trying to build a reliability score.
5-day median matches what I see, but 6 no-replies out of 37 is brutal. Did you count weekends?
Weekends count. Ghosting after 10 days is the signal.
I count weekends too, and day 10 is my cutoff. Started noting auto-reply-to-human gaps in a spreadsheet; three vendors looked fine until I did that. Ghosting is the loudest column now.
Six of thirty-seven stings. Four of twenty-nine went quiet, but two of those came back after two weeks with a generic template. Those get their own bucket in the log: late human, not ghost.
Mass spec mismatches I treat as a yellow flag, not a red one. One of my logged pages had the wrong adduct listed; another was a typo in the parent ion. Still, three in a row would make me pause.
For a reliability score, I'd weight reply time and COA completeness first, ion match second, purity third. A 99% number means little if nobody answers an email. My current sheet is just columns and gut notes.
Ten days is my cutoff too. Weekends count.
92.1% on HPLC made me raise an eyebrow, but method and integration can explain a lot. I note it, then wait for the mass spec. Silence still outranks a low purity number in my book.
I ran a 9-week version of this with 18 suppliers, and my weird outlier was the opposite: a 2-hour human reply with a COA dated 14 months earlier and a lot number that didn’t match anything on their site, while the slowest one (8 days) sent a corrected batch-specific PDF after I asked about the mass spec. Tht flipped my scoring—I
Ha, my outlier was instant reply then total ghosting.
Ha, my outlier was instant reply then total ghosting—same vibe. Mismatched mass spec would be my nope.
One thing I’d split out: auto-reply vs first substantive human reply. My 11-week log (23 vendors, 71 emails) has a 4.1-day median to any human, but only 12 ever answered a lot-specific question; median there was 9 days.
Also, 4 of my mass-spec/HPLC mismatches were recycled old COAs posted against newer lots. Are your three mismatches on the same lot as the purity number, or older PDFs? That distinction changed my red-flag rate from 17% to 4%.
Fellow plateau prisoner here. One thing I started tracking: whether the reply included a lot number and test date that actually matched the posted COA.
Vendors who did replied in ~2 days; generic “we’ll look into it” ones dragged past 10. Also, every never-replier had a COA page last updated >6 months ago—maybe coincidence. Did you split your 214 points by COA recency? Curious if stale docs predict silence better than reply time does.
Login to reply to this thread.
Login / Sign up