10

Big Blue IBM seemingly since forever is using a fixed numbering scheme for all their parts, where every little thing, dingus and gizmo is labeled with the so-called IBM SKU#. It's always in the form:

digit digit letter digit digit digit digit

For example, SKU# 06H8935 is a "front door bezel with keylock", SKU# 92F0406 is a specific "540MB SCSI hard-drive", and so forth. Just check the detailed listing on this x3650 server. Also, on much older equipment, like a line printer model 2390, a cable carries SKU# 92X2785.

There seems to be a unique number issuing system in place at IBM, as all numbers are unique. Also, they seem to be issued completely random as even some research and listing on my side didn't unearth any patterns or product-category-SKU relations. Obvious is that the letter is used to expand the amount of numbers within the fixed 7 position scheme, by using a Base26 element (didn't check if it's actually all 26 letters). And with IBM being IBM, it could be possible they did a field experiment with humans where they tested which number letter combination could be memorized best with fewest errors. Or there's a check digit/letter hidden somewhere.

I'm unsure if this is more of a superuser.com type question - but I'm posting it here as I think it's more related to computer history. I'm moving it ASAP if comments tell me to. And also, I know that the idea of assuming there's an interesting story behind these SKUs lingers on the side of numerology - but if there is something fascinating to tell, it would be a shame to not bring up the question here, for everyone. Right?

Improve this question
7
  • 7
    Another possibility is that the numbering scheme is segmented in order to avoid having a central authority/bottleneck for assigning all SKUs. For example, a prefix ("06H",) assigned to some subdivision, and that subdivision can assign suffixes ("8935") as they wish. Similar to how MAC addresses work. The possibilities are endless. Fun question, I look forward to answers!
    – Wayne Conrad
    Commented Jun 2, 2022 at 17:51
  • 2
    Alpha-numeric codes like this are naturally broken into segments mentally without the use of dashes whereas an unbroken long number can be unwieldly.
    – Brian
    Commented Jun 2, 2022 at 17:56
  • The notion of the letter as (mental or per-subdivision) separator is interesting! I do hope someone, possibly from inside IBM, can shed some light.
    – Micropolis
    Commented Jun 2, 2022 at 19:08
  • 3
    IBM employee ID is digit letter digit digit digit digit. That means there's only ten possible part numbers per employee!!
    – Solomon Slow
    Commented Jun 3, 2022 at 0:02
  • 1
    There's more than IBM involved now, BTW. Lenovo (formerly IBM Thinkpads and other computers) and Lexmark (formerly IBM printers) use the same scheme to this day! (Lenovo example) (Lexmark example)
    – davidbak
    Commented Jun 3, 2022 at 15:12

1 Answer 1

Reset to default
6

Caveat: I'm not into IBM logistics, nor a blue shirt at all. This is what I learned many, many years ago when working with IBM field engineers in the 70s/80s.

  • An SKU is a Stock Keeping Unit.

  • It's the primary key of IBM's stock management system.

  • In (modern day) data base terms it's a Unique (*1), Surrogate Key

  • It's used to identify an items in stock

  • It is only valid during lifetime / time of availability.

  • Its 2-1-4 structure is used for pronunciation.

  • It's purely for logistics purpose.

  • It's otherwise meaningless.

  • New parts to be stocked get new numbers assigned in sequence .

  • Numbers are added when a new product gets ready to be stocked. That is:

    • The product is approved for sale
    • The design process is done
    • The serviceable units / components are defined

  • Numbers have a state associated with 'a date' and 'defined as' (Names mine, can't remember IBM's terms)

    • Upcoming (to be introduced)
    • Alive (active part)
    • Last Order (will be phased out)
    • Phased out (may have different sub states)
    • Gone (no longer visible in ordinary stock handling)

  • IIRC there is an 8th key value (*2) that gets incremented every time an SKU gets reused. (*1)

I'm really sorry that there is no numerology, but pure usability. The kind of usability a stock management (and their databases) needs to function. This was quite common in the early days of stock keeping, and who, if not IBM, the master of punch card management, would use it exactly that way?

*1 - Unique only over its time of availability. The 8th value (*2) makes it unique past that - aka when the "official" SKU part gets reused.

*2 - The 8th value was originally a single character

Improve this answer
5
  • Takeaway no. 1: "the 2-1-4 structure is used for pronunciation", takeaway no 2.: there is an 8th number!? Never saw that. And: if SKUs are issued based on item state, then all SKUs I've ever encountered were "alive" parts. This might explain why the RAID adapter in above's x3650 example has 3 SKUs, one even being a field-replaceable unit (FRU). Also, was hoping for more story behind prefixes etc. than "issued when it arrived at the warehouse" but combining this with Wayne's notion of subdivisions, it could be different warehouses / supply chains.
    – Micropolis
    Commented Jun 3, 2022 at 13:47
  • Leaving this question open for now while hoping for more anecdotal background on these numbers. But cheers Raffzahn, already!
    – Micropolis
    Commented Jun 3, 2022 at 13:49
  • 2
    @Micropolis a) The 8th number is never visible. it's kind of a generation counter, it's incremented by 'one' whenever the number is reused. It will never be visible, and there can only one part with a given 7 digit number in the system at any given time (call it a generation number). c) different SKU may as well be due different exchangeable items - like compatible adaptors from different vendors. d) the idea of subdivisions is rather unlikely. IBM always operated its logistics as a centralized service. If at all, it may have been so in the very early days of mechanical processing (pre-1950).
    – Raffzahn
    Commented Jun 3, 2022 at 14:44
  • 2
    @Micropolis The SKU is a strict logistics related term. in design and production other numbers were used as well. An Example for a 'meaningful' system would have been the Sachnummer (Item number) at Siemens. a company of similar size to IBM (if not larger back then) To be meaningful it needed TWENTYSIX positions (including a leading 'S' and 4 hyphens) It was for parts much structured like you may imagine, denoting department, device, part and alike. And while being unique to identify about anything from books to bolts, for practical (field) usage only a ...
    – Raffzahn
    Commented Jun 3, 2022 at 14:49
  • 2
    @Micropolis ... 6 to 8 digit (depending on department and type) section would be used by engineers to order parts. Not much different from IBM's SKU. just less stringent. I guess the difference is about IBM being much more into data processing at the time either system was developed - they already knew that a unique primary key without any meaning or meaningful structure is way more powerful than anything else - so why using anything else? (Also, My writing is meant as an anecdotal note, not an authoritative answer.)
    – Raffzahn
    Commented Jun 3, 2022 at 14:51

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .