The notice arrives by email and looks like administrative noise. A part number you recognise, a last-order date, a last-shipment date. If the product is still selling, you have just been given a deadline.
The instinct is to buy stock immediately. Sometimes that is right. Often it is the most expensive of the three available options, chosen because it is the one that requires no analysis.
Start with the number that decides everything
Before evaluating any option, work out how many more units you will build. Not the optimistic forecast — the number you would defend.
Everything follows from it. A product with eight months of production left and one with six years left are not the same problem, and the correct answer differs completely between them. Teams routinely skip this step and then argue about solutions without ever agreeing on the question.
Option one: last-time buy
Buy enough of the remaining stock to cover the rest of production.
This is right when the remaining life is short, the part is cheap, and it stores well. It is a bad answer when any of those are untrue. A last-time buy converts working capital into inventory, and inventory carries cost, risk and — for anything with electrolytics or moisture sensitivity — a shelf life of its own.
Size it against the forecast you actually defended, then add a yield allowance. Do not add “a bit extra to be safe”: that is how a nine-month problem becomes five years of stock nobody will write off.
Option two: a qualified replacement
Find a different part that does the same job.
This is the best answer more often than it gets chosen, mostly because it requires engineering time that nobody has budgeted. Cost it honestly: reviewing the alternate, checking the footprint and the control interface, building samples, and re-testing whatever the change touches.
A replacement in the same footprint with the same interface may take a day. One that changes the pinout, the package or the register map is a board revision wearing a disguise — price it as such.
Option three: designing the dependency out
Sometimes the honest answer is that the part should never have been load-bearing. If a single obsolete component is holding a product hostage, and that product has years of life ahead, a targeted redesign of that section can be cheaper across the remaining life than either alternative.
This is also the moment to fix the reason it happened. A board that depends on one irreplaceable part usually has more than one.
Comparing them properly
Put all three on the same page, costed across the remaining production run:
- Last-time buy — parts, plus carrying cost, plus the risk you overbought.
- Replacement — engineering time, samples, re-testing, documentation.
- Redesign — the above, plus layout, fabrication and a longer validation cycle.
Written out this way the answer is usually obvious, and it is frequently not the one the room assumed at the start.
The part that is easy to miss
An end-of-life notice is information about a supplier’s roadmap, not just one part. If a manufacturer is winding down a family, other lines in your bill of materials may be next.
Reviewing the whole bill of materials at that moment costs an afternoon. It is the cheapest hour of obsolescence work you will ever do, because it is the only one you get to do before the deadline exists.