Not recommended for new designs (NRND) is the manufacturer telling you, politely, that a part's days are numbered. Production continues for existing customers, for now, but the roadmap has moved on, and one day an end-of-life notice follows. Nobody sends you a letter about it. The flag sits quietly in a datasheet header or a distributor listing, waiting to be noticed.
We see NRND parts on founder bills of materials (BOM) constantly, and there is a structural reason. Prototypes get built from what is available, affordable and well documented, which usually means the modules hobbyist suppliers sell in volume. Those are, almost by definition, mature parts. Mature parts are the ones closest to retirement. A recent client audit turned up two NRND components on one parts list: the inertial sensor and the satellite positioning (GNSS) receiver, the two parts the entire product depended on. The design worked beautifully. It was also built on borrowed time.
Why this matters more than it seems
A part going end-of-life after you have tooled, certified and launched is one of the most expensive routine events in hardware. The redesign itself is the small cost. The requalification is the large one: a new component can mean new firmware, new testing, and in the worst case a new certification run, because the certificate you paid for describes the board you no longer build. Published research on diminishing manufacturing sources and material shortages (DMSMS) puts the cost of resolving an obsolete part at the design stage against resolving it after production at a ratio of roughly two hundred to one. The same problem, priced at two different moments.
How to check your own bill of materials in an afternoon
Work down the parts list and look up every active component, meaning every chip and module with a part number. Three places tell the truth. The manufacturer's own product page states lifecycle status directly; look for the words NRND, "last time buy" or "end of life," and check the product change notification (PCN) archive while you are there. Large distributors flag lifecycle status on the listing itself. And component search engines aggregate the flags across sources, which makes a whole-BOM sweep fast. Write the status next to each line. An hour of tedium now against a redesign later is the best trade in hardware.
What to do with what you find
The answer depends entirely on where you are in the journey, which is the whole point of checking early.
Before design freeze, an NRND part is a footnote. You pick the successor, the schematic changes a line, and nothing has been lost. This is the cheapest possible moment to know.
The one trap at this stage is the false drop-in. A replacement must match more than the footprint and the interface. It must match the performance your product actually depends on. A sensor family successor can share the pinout and quietly halve the output rate that your application requires. Read the successor's datasheet against your real requirements, not against the old part's headline.
After tooling but before certification, a swap still costs board revision and firmware work, but the certificate is not yet at stake. Move quickly and the damage stays contained.
After certification, you are choosing between requalifying and buying time. Buying time is legitimate: a last-time buy, a calculated stock of the dying part covering your production forecast, is a standard and sometimes correct move. It needs real arithmetic on volumes, shelf life and cash, and it needs a plan for what happens when the stock runs out, because it will.
And in every case: prefer parts with a published longevity commitment where the category offers one, and note second sources where they genuinely exist. Some component categories have true alternatives; many do not, and knowing which is which is half the value of the exercise.
Where this fits in the bigger picture
Lifecycle status is one of the checks inside a production readiness audit, alongside manufacturability, cost drivers and the certification path. It is also a good example of why automated design checkers miss what matters: a rule engine sees a valid footprint and a connected net, and has no opinion about whether the part will exist in eighteen months. Someone has to ask. Our audit asks, in writing, before any money reaches tooling.
Common questions
What does NRND mean on a component?
Not recommended for new designs (NRND) is the manufacturer telling you, politely, that a part's days are numbered. Production continues for existing customers, for now, but the roadmap has moved on, and one day an end of life notice follows. The flag usually sits quietly in a datasheet header or a distributor listing rather than arriving as a notification.
How do I check my bill of materials for NRND parts?
Work down the parts list and look up every active component, meaning every chip and module with a part number. Three places tell the truth: the manufacturer's own product page, which states lifecycle status and holds the product change notification archive, large distributors, which flag lifecycle status on the listing itself, and component search engines, which aggregate the flags across sources. Write the status next to each line.
What should I do if an NRND part is found after certification?
After certification you are choosing between requalifying and buying time. Buying time is legitimate: a last time buy, a calculated stock of the dying part covering your production forecast, is a standard and sometimes correct move. It needs real arithmetic on volumes, shelf life and cash, and a plan for what happens when the stock runs out, because it will.
Lifecycle status is one line in a much longer check.
The Production Readiness Audit puts an engineer on your files and returns a written manufacturability report, the cost drivers, the required changes and a clear verdict, at a fixed price agreed before we start.
Start with your product