Thursday, 4:47 PM — The Phone Rang
March 2024. My cell phone buzzed. The caller ID said Todd Pepsi.
Todd runs procurement at a mid-sized network integrator. We'd been working together for about two years. He's direct — usually starts talking before I even finish saying hello. This time, he paused. Two full seconds of silence, then: "I need a favor. And I need it fast."
Here's the situation. Todd's company was doing a campus network upgrade for a manufacturing client — roughly 200 Cisco switches and routers across three buildings. The project had already lost three weeks to internal debates over the architecture: the client's IT team kept going back and forth on networks vs Cisco gear. Mixed-vendor to save cost? Or all-Cisco for uniformity? By the time they settled, there was barely two weeks left for hardware testing and commissioning.
The bottleneck sat on a batch of power boards. Those boards used Kemet T520 series tantalum capacitors, paired with a small run of Kemet inductors. Todd had ordered through another distributor. That distributor didn't mention a potential delay until day four. Standard lead time was 3-4 weeks. A delay meant the entire network commissioning would slip — and the client's acceptance deadline was already written into the contract.
"If we don't have those parts by Saturday morning, the whole rollout slips a week. And I don't have a week."
I looked at the clock. He had less than 40 hours to get those parts into his hands.
You might ask: why not call Kemet directly? Todd had tried. Kemet's direct channel has MOQ requirements and standard lead times that don't bend for small emergency orders. The volume he needed — a few hundred T520 units plus a batch of inductors — was small enough that authorized distributor inventory was his only realistic shot.
Step One: Confirm "Can We Even Do This?"
After 11 years of handling emergency orders, the first rule I've learned is: don't say yes until you've checked. Sounds obvious. You'd be surprised how many people skip it.
I asked Todd to send the exact part numbers and quantities. After we hung up, I pulled up our inventory system. The Kemet T520 line alone has dozens of variants — voltage ratings, capacitance values, case sizes (A, B, C, D, X and so on). Todd needed the T520 series in a B-case package with specific capacitance and voltage specs. He also listed the Kemet inductor part numbers.
The system showed partial inventory of the T520s in our US warehouse — not enough to fill the order. The Kemet inductors, though, were fully stocked. That was the easy part.
I ran the timing. If we pulled the T520s from the US warehouse and topped up from our Asia warehouse, we could theoretically have the full quantity by Friday afternoon. But there were variables: cross-warehouse transfer logistics, customs clearance, outbound processing. Three things that could each eat a few hours I didn't have.
I called Todd back. I said: "I can do it. But I'm going to be straight with you. The T520 portion has to come from two warehouses. If everything clears by 3 PM Friday and ships overnight, you'll have it Saturday morning. But if anything slips — inventory doesn't match, or the courier has a delay — I can't promise it."
Silence. Then: "I need a clear answer."
I said: "I'll give you 80% confidence. The other 20% — you can start preparing a Plan B right now, or we can take the gamble together."
This is the part where a lot of people in my position would have just said "no problem." I've done that before. It doesn't end well.
Friday, 8 AM — The Turn
I got to the office early. First thing: check the US warehouse outbound status. The system said "processing." That felt wrong. A transfer order placed the night before should already be in the packing stage by morning.
I called the warehouse. Waited 12 minutes on hold. The answer made my stomach drop: "The T520 batch — system says we have it, but we only found half on the shelf. The rest might be pre-allocated to an earlier order. The system didn't update."
I'd broken my own rule. Two years ago, I trusted system inventory on an emergency order. System said we had it. We didn't. The client waited six days, then took their business elsewhere. After that, I made myself a rule: for emergency orders, always call the warehouse. No exceptions. Thursday night was too rushed. I skipped the call. I still kick myself for that one.
I asked: "When will the other half arrive?" The warehouse said around 2 PM, transferred from another location. I did the math — if it arrived at 2, packed and shipped by 3, overnight delivery, Todd would have it Saturday morning. Still a shot. But I couldn't be sure it would actually arrive at 2. Last month, a transfer was supposed to arrive in the afternoon and didn't show up until the next morning. That one nearly cost me a client in industrial controls — their line was down for a full day. They recovered, but now they double-confirm lead times on every order.
I texted Todd: "Change of plan. US warehouse inventory didn't match — the other half arrives at 2 PM. I'm watching it. You'll hear from me the second I know anything."
He replied fast: "What if it doesn't arrive by 2?"
I stared at the screen. The upside of waiting? If the transfer arrived on time, everything worked. The downside? If it didn't, Todd lost his window. I typed: "If it's not here by 2:30, start your Plan B. But I'll tell you exactly where things stand. No guessing games."
Then I did something I should have done the night before: I confirmed the Asia warehouse stock on the T520s. On paper, it looked fine. On the phone, the warehouse manager said: "We have the quantity, but one reel has a different date code. Should still work, but I'm flagging it." Different date code. Same spec. Probably fine. But "probably" wasn't a word I wanted to say to Todd at 2:30 in the afternoon.
I called Todd. "One more thing. The backup stock has a different date code. Same spec, but it's not the same production run. If you're doing incoming inspection, flag it."
He said: "Will that matter for our testing?"
I said: "Honestly? I don't know. It could be nothing. It could be something. If you have a test bench, check the first few units before you populate the whole board."
That was the moment I realized I probably should have just told him to find another source. The complexity was piling up. But at that point, he had no other options within 48 hours. And I did have a path, even if it was messy.
2:17 PM — It Arrived
At 2:17 PM, the warehouse called. The transfer was in. I told them to pack it immediately and push it through the fastest process they had. At 3:05 PM, the tracking number came through. USPS Priority Mail Express — guaranteed by 10 AM Saturday for most US addresses if shipped before the cutoff. We made the cutoff by 25 minutes.
I sent Todd the tracking number: "Shipped. Arrives tomorrow morning. Have someone at receiving by 10 AM."
Saturday, 9:40 AM. Todd sent a photo. Box arrived. Seal intact. He wrote: "Thanks. That was close."
I replied: "Don't cut it that close next time."
He sent back a smiley face. There's something satisfying about pulling off a run like that — after all the stress and the date-code phone call and the 12-minute hold with the warehouse, seeing the box arrive and the parts work. That's the payoff.
What I Took Away From This
I went back through this order afterward. A few things stuck with me.
One: the first rule of emergency orders is "ask first, promise second." I've watched people in this business grab orders by saying "no problem" before they check anything. Then when they can't deliver, the client doesn't come back. When I told Todd "80% confidence," that wasn't me hedging. It was giving him a real number to work with. And that early honesty saved me — when the inventory issue came up, he was already mentally prepared. He wasn't blindsided.
Two: system data always needs a discount. I learned this two years ago and still got lazy. Now I've added a hard rule to my emergency order process: any cross-warehouse transfer requires a phone confirmation of physical stock. No exceptions. I almost got burned on this again.
Three: being honest about limitations keeps clients better than overpromising. If I'd said "100% no problem," Todd might have felt more confident placing the order. But if something had gone wrong — and it nearly did — the trust would have been gone. Instead, I told him "there's a 20% risk, your call." Later he told me: "That line right there? That's why I'm calling you next time."
Now — this approach isn't universal. If you need high-volume Kemet T520s or Kemet inductors — say, tens of thousands of units — the emergency sourcing window gets much smaller. Authorized distributor inventory depth has limits. In that case, standard manufacturer lead times are probably more realistic. And if you need delivery timed to the hour, emergency sourcing will disappoint you — we can tell you "roughly what time it'll arrive," not "2:37 PM exactly."
There's also a category of client I'd quietly steer away from the emergency route: anyone whose production line is already running and just needs maintenance parts. In those cases, paying a 15-20% premium for emergency sourcing usually doesn't make sense unless the line is actually down. Wait for standard lead time. Your CFO will thank you.
As for Todd — the project passed acceptance without a hitch. He called again last month. This time he gave me three weeks' notice. On the phone he said: "No rush this time."
I said: "Good. I'd like to sleep at night."
He laughed and hung up. That's the version of this job I prefer.