I’ve spent over a decade in medical device supply chain. Class II and Class III devices. Domestic and overseas sourcing. In that time, I’ve watched a lot of founders run supplier qualification for the first time. Most of them learn the hard way.

The patterns are consistent enough that I can name five principles. They’re the framework I use when I help founders today.

1. Stage your supplier network like your product roadmap #

Most founders treat supplier selection as a one-time decision. Pick a supplier, qualify them, move on. The problem is that your needs change as you progress. What works for a prototype run often fails at production volume.

The prototype supplier who turns samples around in a week may not have the capacity or quality systems for clinical batches. The clinical supplier with full ISO 13485 traceability may be too slow and expensive for commercial scale. These aren’t failures. They’re stage transitions you didn’t plan for.

Plan supplier transitions on your roadmap, right next to your design milestones. When you map out prototype build, design freeze, clinical, and commercial, note which suppliers come with you and which graduate or get replaced. Make those transitions visible before they become crises.

2. Map the full chain before qualification starts #

When you send a part out for qualification, you’re not qualifying one supplier. You’re qualifying every tier behind them.

A printed assembly has a sub-supplier making the raw material, another doing the coating, another doing the heat treat. Each tier adds documentation, controls, and time. If you haven’t mapped the full chain, your timeline estimate is a guess.

During RFQ, ask for a sub-supplier map. Where does the raw material come from? Who does the finishing, the inspection, the sterilization? If a supplier won’t or can’t tell you, that’s a signal, not a paperwork problem you chase later.

3. Inspect the process, not just the product #

First article inspection tells you today’s parts are good. It doesn’t tell you tomorrow’s parts will be.

The supplier’s process is what determines consistency. If you don’t have a production flow diagram (step by step, including outsourced operations), you’re qualifying outputs from a black box. When that black box drifts, you’ll find out after a bad batch, not before.

Make the production flow diagram a standard RFQ requirement. No flow diagram, no qualification. Treat it like an audit certificate, not a nice-to-have.

4. Optimize for change-tolerance, not spec compliance #

Here’s a truth that surprises most first-time founders: your design will change many times before commercial. Not small tweaks. Real changes in geometry, materials, tolerances.

A supplier who hits today’s spec perfectly but freezes when you ask for a change is more expensive than one who’s slightly behind on the initial spec but adapts fast. Spec compliance is a static check. Change-tolerance is a dynamic capability. Most evaluation frameworks only test the first.

When you evaluate a supplier, pay attention to how they handle ambiguity. Ask about a time a customer changed direction mid-stream. How did they respond? If they can’t give you a real example, that’s information.

5. Preserve optionality until you must commit #

Single-source is a commercial-stage commitment, not a default. But most founders drift into it without deciding to.

The cost of keeping a viable backup warm is small. A few thousand in samples, a few hours of capability review, maintaining the relationship with one alternative supplier. The cost of qualifying a backup in a panic, for a specialty component, is six to twelve months.

Identify a viable alternative for every critical component before you need one. You don’t have to qualify them today. You have to know they exist, they’re capable, and they’d take your call when something breaks.

These five principles are the starting point. Underneath them is one frame that ties it all together: supplier selection has two layers. The technical layer (can they make the part?) you can buy through credentials and audits. The adaptive layer (can we work together as things change?) you have to build through interaction. Most founders only evaluate the first.

Got a concrete sourcing problem? Book a 1:1 with me — $500 (30 min) or $1,000 (60 min). Real answers for founders who need a specific decision made, not general advice.

If you don’t have the bandwidth to run the process yourself, the Sourcing Sprint runs it for you. Four weeks, one or two qualified suppliers and a recommendation memo.

This article appears in the Agile Supplier Field Guide — a strategic guide for pre-Series A medtech founders.