A few years ago I worked on a wellness device: a small object meant to help people relax through patterned vibration instead of sound or light. Simple concept, simple bill of materials on paper. One part carried most of the product’s identity: a compact haptic motor tuned to produce a vibration people would actually find soothing rather than buzzy or mechanical.
That motor was easy to source when we picked it. A few months after we’d finished the design, it wasn’t. The distributor’s answer went from “in stock” to “call for lead time” almost overnight, the kind of shift that has nothing to do with anything we did and everything to do with how thin that corner of the component market actually is.
So we did what you’re supposed to do: we asked for an equivalent. The distributor came back with a part that matched on paper. Same footprint, same voltage range, same rough current draw. By every box a datasheet comparison checks, it was a drop-in replacement. It was also five times the price.
That gap is the real problem, and it’s not really about this one motor. Haptic actuators in the size and quality range wellness and wearable products need sit in a much thinner slice of the supply chain than, say, a resistor or a generic MOSFET. Fewer manufacturers make them. Fewer distributors carry meaningful stock. Pricing on the “equivalent” part reflects how tight that specific niche happens to be that quarter, not the fact that the two components look interchangeable in a comparison table. A datasheet tells you two parts can fit the same footprint. It says nothing about whether either vendor can actually deliver volume at a price your BOM can absorb this year.
The trade-off nobody wants to hear is that there’s no clean rule that catches this ahead of time. I’d love to hand a client a checklist: dual-source anything with a lead time over eight weeks, avoid single-vendor actuators, flag any part with fewer than three manufacturers. I’ve tried versions of that, and they don’t hold up project to project. A rule tight enough to catch the haptic motor problem also flags parts that never actually cause trouble, and a rule loose enough to leave those alone misses the next surprise, which usually isn’t a haptic motor at all, it’s a connector, or a battery cell format, or a niche RF module. Judgment beats a checklist here, evaluated part by part: how specialized is it, how many real alternates exist, how exposed is the design to this one supplier’s roadmap.
There’s a timing dimension that makes this worse than a line-item cost surprise. By the time the shortage hit, we were already deep into calibration tied to that motor’s specific vibration characteristics, the kind of tuning work that’s genuinely hard to redo once it’s closed out. Swapping the part wasn’t just accepting a 5x cost. It risked reopening testing we’d already signed off on. That’s the part that stings: the cost isn’t the new part price, it’s the rework the timing forces on you.
That timing problem connects to something I tell clients that surprises most of them: before you commit to volume, and before you commit to formal certification, you need real market feedback, not internal conviction that the product is right. It’s a genuine grey area. Certification and tooling for volume both want a frozen design, and a frozen design is exactly what turns a supplier hiccup into a re-certification problem instead of a shrug. Committing to that freeze before you know the product will actually sell is a bet you’re placing twice: once on demand, once on your BOM staying stable long enough to matter.
The practical path is less satisfying than a rule but it’s what actually holds up. Early in a design, walk the BOM and flag the parts that are genuinely thin in the market: specialized actuators, radios, anything from a single manufacturer with no drop-in family. For those specific parts, get a second source at least loosely qualified, even just a candidate you’ve confirmed fits electrically, before you need it under pressure. Separately, resist locking certification and volume commitments to internal confidence alone. Run a real pilot batch, get it in front of actual users, and let that feedback tell you when the design is stable enough to freeze. The two problems compound each other: an unqualified single-source part is a manageable risk on a design that’s still open, and a serious one on a design you’ve already frozen for certification.
I don’t think there’s a version of this where you avoid every surprise. Components go end-of-life, distributors run out, and “equivalent” will keep meaning different things to an engineer and to a supply chain. What you can control is not being fully committed, on cost or on timing, before you’ve tested both the part’s real availability and the product’s real market fit. That’s judgment, not process, and it’s uncomfortable for anyone who wants a spec sheet answer. It’s still the honest one.