Part two of a three-part series on lean versus redundancy in hardware. Part one was about your team. This one is about your supply chain.
The Shortage Doesn't Announce Itself
You don't find out your connector is single-sourced when you design it in. You find out when your contract manufacturer emails you three weeks before a production run to say supply is gone.
That's not hypothetical. It happened on a product I was managing. We'd designed around a specific connector, sent the package to our CM, and got the call: not enough supply. Now the timeline that took months to build is hostage to a part nobody thought twice about.
Every hardware team carries this same exposure. The only question is whether you find it before your CM does, and whether you actually know what your CM is sourcing in the first place.
Supply Chain Means More Than Parts
When people say "supply chain risk," they usually mean the parts on the BOM. That's only half of it. Where those parts get made, and where your product gets assembled, carries the same risk profile.
The full version of this strategy is geographic diversification: main manufacturing in China, a second line in the US, a third in Mexico, so a tariff spike, a factory shutdown, or a border closure in one region doesn't take down your entire production run. Companies that can afford it are already doing this.
You probably can't, not yet. Running parallel relationships with contract manufacturers on two continents is expensive and slow to set up, and it's not where a startup's redundancy budget should go first.
Start With What Would Actually Stop You
Before you think about manufacturing footprint, look at your BOM and ask a narrower question: which parts, if they disappeared tomorrow, would actually halt the product.
Not everything on the BOM carries that risk. Sheet metal, copper pipe, standard steel stock: these are commodities. Any machine shop can source equivalents, and swapping one supplier for another changes nothing about your design.
The parts that matter are the specific ones: a particular microcontroller, a specific connector, anything where you designed around one exact part number instead of a category of parts. Those are the ones a CM email can turn into a schedule slip.
Go through your BOM and flag those. That's the whole first step. You're not building a dual-sourcing program yet, you're just finding out where your exposure actually lives.
Alternates Aren't Free, But They're Cheap Compared to a Delay
Once you've flagged the risky parts, the next move is finding an alternate for each one, and accepting that the alternate probably won't be a drop-in replacement.
In my case, the fix for the connector meant a design variance. Connectors carry their own footprint: a specific diameter, keying, mounting constraints into the enclosure. Switching connectors touches the design around it. It's not a large change, but it's not nothing either.
That's the tradeoff. A minor design variance now costs you far less than a production delay discovered the week you needed to ship.
Where This Lives in Your BOM
Once you've got an alternate, you need somewhere to record it that isn't a Slack message to yourself.
There are two ways to handle it in a PLM. You can carry the two connectors as genuinely separate parts, each with its own BOM line and its own place in the product structure. Or, if you want to keep it simpler, you assign one internal part number to the connector and track the primary and alternate as variants under it, with a note on what changes elsewhere in the BOM if you switch: which enclosure revision it pairs with, which other parts move with it.
Either way, the point is the same: the alternate exists somewhere other than your memory, and anyone touching that BOM six months from now can see it without having lived through the shortage themselves. A PLM system is where that record actually survives; a Slack thread from the last shortage is not.
The third part of this series covers the same tension from the engineering side: how much redundancy you design into the product itself, versus how lean you build the BOM.
Run lean on your supply chain if you have to, but know which parts you can't afford to. If you want one place to track part alternates alongside your BOM and revision history, see how it works at oroforge.com.
