It started with a standard capacity upgrade
Back in Q2 last year, our company—a mid‑sized telecom equipment distributor based in De Soto, KS—needed to expand our demantra network. We were about to roll out a new service for a customer called Clear Phone Inc., and that required a batch of Infinera-compatible SFP+ and QSFP56 transceivers for our DTN‑X platform. My annual budget for optics was roughly $180,000, and I had six years of invoice‑tracking under my belt. This should have been routine.
The project lead gave me three weeks to source and qualify the modules. I started by reaching out to my usual vendors. One name kept surfacing: David Welch. He runs a regional distributor that specializes in Infinera Nokia telecom solutions—apparently their combo of Infinera gear and Nokia software is getting traction. David called me directly. “We can do the whole lot for 22% less than your current supplier,” he said. The numbers were tempting.
The low‑price bid – and the hesitation
I went back and forth between David’s offer and the established vendor for two weeks. David’s quote was cheaper. Significantly cheaper. But something felt off. Every time I looked at the datasheet for his modules, the specs matched, but I couldn’t shake a mental image of that cheap power supply I’d once bought that blew a capacitor in a month.
That thought brought me to a weird place. I actually pulled out an old multimeter and tested a capacitor from a failed module I’d kept in my drawer (note to self: never throw away SBLs when you have a hunch). If I remember correctly, the reading was 470μF but showed only 300μF. That module? From a bargain‑brand vendor. I learned then that testing a capacitor with a multimeter isn’t just for hobbyists—it’s a litmus test for build quality. The “cheap” vendor had skimped on components. (Ugh.)
The trigger event that changed my mind
While I was still hesitating, a different project hit a snag. We had deployed a batch of supposedly Infinera‑compatible SFP+ modules from Vendor C (not David, another budget option) in a customer’s edge router. After 48 hours, the link started flapping. The field engineer spent a full day troubleshooting—checking fiber, cleaning connectors, swapping SFPs. He eventually found that the module’s voltage regulator was unstable. The capacitor on the regulator rail was underspecced (yes, again). That single failure cost us three man‑days and a lot of client goodwill.
That incident in May 2023 changed how I think about component quality. The “savings” from the cheap modules vanished the moment we had to send a technician on‑site. Total cost of ownership, I reminded myself—not just the sticker price.
Total cost thinking – the real anchor
Here’s the framework I’ve come to rely on, and it’s the same one I use for every Infinera‑related procurement:
Total cost of ownership includes:The lowest quoted price often isn’t the lowest total cost.
- Base product price
- Shipping and handling
- Potential rework or replacement costs (quality issues)
- Downtime impact during failures
- Burned engineering time for troubleshooting
— My own cost‑tracking spreadsheet, updated after every order since 2019
In the Clear Phone project, David’s offer looked good on paper, but when I ran the TCO model—including a conservative 2% failure rate and the cost of a single truck roll—the established vendor’s modules came out 8% cheaper over a three‑year lifecycle. (David’s modules might have been fine, but I couldn’t verify the capacitor brand without cutting one open. And given my luck? Pass.)
The pivot – sticking with quality
So I went with the premium‑priced vendor. The modules arrived on time, installed cleanly, and Clear Phone’s service went live without a hitch. The client was happy. More importantly, our sales team started using that cite as a reference for future deals. “We run fully qualified Infinera‑compatible optics – no flapping, no surprises.” That became a differentiator.
Was the extra 18% on the invoice painful? Yes. (I’m a cost controller – it hurts.) But the moment I saw the client feedback score improve by 15 points on “technical reliability,” I knew it was worth it. Quality perception is brand experience. If you hand a customer a flaky transceiver, they don’t blame the vendor; they blame you. Your brand becomes the cheap one.
The lesson I cycle back to
After that project, I formalized a policy: every new transceiver vendor must provide test data for at least three sample units, including DC‑DC converter ripple and capacitor ESR measurements. My engineers laughed – until they saw how many vendors quietly failed. Now we share those results in quarterly reviews.
I don’t think David Welch’s modules were bad. In fact, I’ve since bought some Infinera‑compatible CFP units from him for a less critical lab setup, and they’ve been fine. But for production? I’ll pay the premium. The certainty of knowing the parts meet specifications—down to the capacitor inside—is worth more than a 20% discount on paper.
Sometimes the best procurement decision isn’t the one that saves the most money today. It’s the one that protects your reputation tomorrow. Period.
