Most card scanners are built for one card, not a hundred.
Most scanners are built for the one-card office moment, not the hundred-card reception. Here are the four design failures that surface at scale, and what gets it right.
The line forms behind you
It is the second hour of the reception. You have a stack of maybe eighty cards held together with a rubber band, and the person in front of you is still talking. You pull out your phone, open the scanner, and point it at the card they just handed you. The app captures it, then asks you to confirm the name. Then the company. Then it spins for a moment while it syncs. Then it asks you to add a tag before you can scan the next one. By the time you look up, the conversation has moved on without you, and there are three more people waiting to hand you cards. So you stop scanning. You shove the rest of the stack in your bag and tell yourself you will deal with it later. You will not deal with it later. By Friday, those eighty cards are anonymous strangers.
The office-desk assumption
Here is the thing almost every card scanner gets wrong, and it is not a bug. It is a design assumption. These apps are built for the office-desk moment: one card, plenty of time, you want the data clean and filed correctly before you move on. At a desk, confirming the name and adding a tag is good hygiene. It feels thorough. The whole interaction is tuned for the case where you have exactly one card and nothing else competing for your attention. The conference floor is the opposite case. You have a hundred cards, two hours, and a live conversation in front of you. The constraint is not accuracy, it is throughput, and a tool tuned for one-at-a-time hygiene collapses the moment the cards start arriving faster than you can confirm them. Call it the office-desk assumption: the app was designed for the moment you are least likely to be in when you actually need it.
Accuracy is the wrong thing to optimize
Walk through how this category markets itself and you will see the same headline metric everywhere: capture accuracy. How cleanly it reads the name, how well it parses the title, how few characters it gets wrong. Accuracy is a real thing and it matters, but it is the metric for the desk, not the floor. At conference scale, a scanner that reads cards at 99 percent accuracy but makes you confirm each one is slower and more frustrating than one that reads at 95 percent and lets you fire through the stack. The reader who has actually worked a busy reception knows this in their hands: the bottleneck was never the reading, it was everything the app made you do between cards. Optimizing accuracy while ignoring throughput is how a whole category convinced itself it was solving the conference problem when it was really still solving the desk problem. (We made a version of this argument before, about why reading the card was only ever the easy part: see the piece on OCR being a feature, not the product.)
The four shapes a scanner fails in
Look across the category and the failures are not random. They cluster into four shapes, each one a design choice that is reasonable at a desk and ruinous at volume. First, one-by-one tap-and-confirm: every card requires a manual confirmation tap before the next scan, so your effective speed is capped not by the camera but by your thumb. Second, cloud-sync-blocking-next-scan: the app holds the camera while it pushes the last card to a server, which means a weak conference wifi signal or a dead cell zone freezes the whole flow. Third, per-card editing prompts: the app surfaces an edit screen after each capture to let you fix the parse, which is helpful for one card and a wall when there are ninety behind it. Fourth, forced-tag entry on each scan: the app makes you categorize or tag the contact before it will move on, turning a two-second capture into a ten-second decision you do not have time to make in a live conversation. Any one of these stalls you. Most apps ship two or three of them stacked together.
What it looks like when all four stack
Picture the realistic version. You are at a sponsor reception, drink in one hand, phone in the other. Someone hands you a card and keeps talking. You scan. The app asks you to confirm the name (shape one). You tap to confirm and it stalls, trying to sync over the venue wifi that everyone in the room is also fighting for (shape two). It finally moves, then opens an edit screen because it is not sure about the title line (shape three). You skip the edit, and before it lets you scan the next person it asks which list to file this contact under (shape four). Four interactions, maybe fifteen seconds, for one card, while a real human is mid-sentence in front of you. Do that math against eighty cards and you understand exactly why the stack ends up in the bag. The app did not fail because it could not read the cards. It failed because it was never designed to get out of your way.
Throughput is the metric that matters
If the thesis holds, the fix is not better reading. It is a different design posture for the moment of volume. A scanner that respects conference scale makes four choices that mirror the four failures. It captures continuously, so you can fire through a stack without a confirm tap between each one. It works offline first and syncs later, so a dead signal never blocks the next scan. It defers editing, parking the parse for review when you are back at the hotel instead of demanding it on the floor. And it makes categorization optional and after-the-fact, so tagging is a thing you do later, in bulk, not a toll you pay per card. None of this is exotic. It is just the recognition that the two-hour reception is a different job than the quiet desk, and the tool should change shape for it. There are a handful of ways to build that. Met's Event Mode is one of them.
How Event Mode handles the floor
Met's Event Mode is the version of this we built, because the conference floor is the moment Met was made for. You turn it on when you walk into the room and it keeps capturing while you keep talking: no confirm tap between cards, no sync stall, no edit prompt mid-conversation, no forced tag. The cards land, you stay in the conversation, and the cleanup waits until you are ready for it. Storage is iCloud-only, so the contacts you capture stay yours and are not parked on someone else's server or sold on. It is not the only way to respect the throughput constraint, but it is built around that constraint from the start instead of bolting volume onto a desk tool. If you have ever watched a stack of cards go cold in your bag because the app made you stop, that is the specific failure Event Mode exists to remove.
Get Met and turn on Event Mode before your next conference. Capture the room without breaking the conversation.
Met was built for the conference floor, not the desk. Event Mode keeps the room captured while you keep talking, stores your contacts in iCloud so they stay yours, and leaves the cleanup for when you are ready. It is the one designed around the moment of volume, not the moment of quiet.
