“But the shipment arrived four days early, so how is the responsiveness score a 31?”
“Because they didn’t log the confirmation in the portal, Mason. If it isn’t in the portal, it didn’t happen. The system flagged them as non-responsive for .”
“They called me twice. I have the logs.”
“Did they call the portal?”
That is the exact moment I realized my procurement software was a hallucination. It was a well-funded, sleekly designed, multi-million dollar hallucination that lived in a world where human intention didn’t exist unless it was formatted as a UTF-8 string in a specific text box. We were sitting in a glass-walled conference room that smelled faintly of dry-erase markers and expensive air filtration, looking at a spreadsheet where a supplier in Busan-the most reliable partner we had-was glowing a violent, pulsative red.
The scorecard didn’t care that the inventory was already being unloaded at the dock. It didn’t care that the account manager, a woman named Ji-won who worked shifts to align with our time zone, had reached out immediately. It only cared that she spoke in her native Korean to a system that was essentially a digital wall. To my software, Ji-won wasn’t a partner; she was a lag time.
I’m currently writing this with a slight tremor in my hands because ago, I accidentally hung up on my boss while trying to explain why our “Tier 1” vendors were all underperforming. I was trying to hit the mute button to sneeze, and instead, I severed the connection. It’s the perfect metaphor for the week. We build these interfaces to connect us, but one wrong tap, one missing translation layer, and the relationship is just… gone. The line goes dead. The scorecard turns red.
A small exclamation point that weighs more than of steel currently sitting in our warehouse.
Metrics are ghosts that haunt the machine. One hyper-specific physical example of this is the way a “low engagement” flag looks on a 4K monitor: a small, pixelated exclamation point that weighs more than the twenty tons of steel currently sitting in our warehouse.
Punishing the Best People
We treat responsiveness as a moral virtue when, in reality, it is often just a technical compatibility check. If I send a message in English and you reply in Korean, and my system cannot “digest” that Korean in real-time, the system records silence. It doesn’t record a bridge being built; it records a gap. In the world of high-stakes procurement, silence is a fireable offense. We are punishing our best people because our tools are linguistically illiterate.
“
“A scorecard is just a map drawn by someone who’s never seen the mountain.”
– Elias, Logistics Lead
My friend Elias, a logistics lead who has spent more time on cargo ships than in office chairs, once told me: “A scorecard is just a map drawn by someone who’s never seen the mountain.” He’s right. We look at the map and say, ‘The mountain is too slow,’ while the mountain is actually standing right in front of us, trying to hand us a bottle of water.
This is where the frustration peaks. You see the dashboard, and you see a vendor failing. You start the “de-risking” process. You look for alternatives. You spend months vetting a new supplier who speaks perfect English but has a higher failure rate on the actual product. You trade competence for “render-ability.” It is a catastrophic trade.
The logic of the scorecard is a closed loop. It assumes that if a communication isn’t captured, it didn’t occur. But in a global economy, the most vital communications happen in the “dark spaces” between platforms. They happen in the quick phone call, the voice note, the frantic explanation given in a language that the CRM doesn’t support. We are effectively deafening ourselves and then complaining that the world is too quiet.
I’ve spent the last decade as a debate coach, and the first thing we teach is the “burden of proof.” If you can’t prove you said it, you didn’t say it. But in international business, there is a secondary, hidden burden: the burden of rendering. It isn’t enough to speak; you must be rendered into the listener’s reality.
When Ji-won called, she was doing her job. She was being “responsive” in every human sense of the word. But because her voice was processed as “unstructured audio” that didn’t match the English-only logic of our live-tracking tool, the tool simply waited. It waited for a trigger it recognized. While it waited, it began to deduct points. It deducted points for her accent. It deducted points for her promptness. It deducted points for her existence.
This is the hidden tax of the monolingual workspace. We aren’t just losing time; we are losing the ability to see who our allies actually are. We are firing the people who save us because they didn’t save us in a way our software could count.
The Series of 0.5s
Standardization is a predatory animal. It eats nuances and spits out averages. We want everything to be a 1 or a 0, an “On Time” or a “Delayed.” But the world is a series of 0.5s. It’s a “The boat is stuck but I found a truck.” It’s a “I’m calling you now to tell you the portal is broken.”
If you can’t absorb those 0.5s, you are flying blind. This is why I’ve become obsessed with tools that don’t force people into a specific linguistic box. We need layers that sit on top of the conversation, not gates that stand in front of it. We need something like
to act as the nervous system for the global office. If the software can hear the Korean, translate the Korean, and log the intent-all while the conversation is actually happening-the scorecard finally starts telling the truth.
Without that real-time bridge, you aren’t managing a supply chain; you’re managing a series of misunderstandings. You’re looking at a screen and getting angry at a person who is currently working their heart out for you.
The 70% Gap: Where companies go to die and trust erodes.
I remember looking at Ji-won’s stats after we finally manually overridden the system. Her “Human Responsiveness” was . Her “System Responsiveness” was . That 70% gap is where companies go to die. It’s the space where trust erodes. It’s the space where a buyer gets a “performance improvement plan” for a vendor who is actually their MVP.
We have to stop let the metrics lie to us. A data point is only as good as the sensor that captured it. If your sensor only picks up English, you are missing of the global manufacturing heart. You are essentially trying to see a sunset through a keyhole.
I think back to that hang-up with my boss. The silence after the call ended was heavy. It was a technical error, a slip of the thumb on a glass screen, but to him, it might have looked like a protest. Or a breakdown. Or a lack of respect. In the absence of a clear, translated signal, we fill the silence with our own worst assumptions.
A red cell in a spreadsheet is often just a ghost of a conversation that the software refused to hear.
The vendor scorecard does the same thing. It fills the “untranslated” space with the word “Lazy.” It fills the “unstructured” space with the word “Unreliable.”
We are currently in a transition period in human history. We have the connectivity of a Type I civilization but the linguistic silos of the ancient world. Our tools are trying to bridge that gap, but they often do it by forcing everyone to speak one language-the language of the software’s creator. That isn’t a bridge; it’s a bottleneck.
We need our workspaces to be as multilingual as the docks and the factories they manage. We need the ability to hear a voice, regardless of the language, and have that voice register as “Action” in the eyes of the machine.
When I finally got back on the call with my boss, I didn’t apologize for hanging up. I told him the truth: “The technology failed to capture my intent.”
He laughed, but I wasn’t joking. It’s the same thing I told the procurement board when they tried to drop the Busan supplier. “They didn’t fail us,” I said, pointing at the glowing red scorecard. “We failed to listen to them. We built a system that rewards the loudest English speaker and punishes the most efficient partner. If we keep doing this, we deserve the supply chain we get.”
The Raw Data of Human Partnership
I’m done with scorecards that require a translator to even read the “Responsiveness” column. I want the truth. I want to hear Ji-won in real-time. I want the metrics to catch up to the reality of the work. Until then, I’m keeping my own notes. I’m looking at the warehouse, not the screen.
Because at the end of the day, you can’t build a business on 1s and 0s if the 1s are lies and the 0s are just voices you didn’t bother to understand. We need to stop treating language as a barrier to be “dealt with” and start treating it as the raw data of human partnership. If your software can’t do that, your software is just a very expensive way to stay lonely in a crowded market.