4 ms·
"WeWork used [a sensor network that tracked everyone throughout the building combined with machine learning] to learn that people drink coffee in the morning so
by mandevil 7y ago
"WeWork used [a sensor network that tracked everyone throughout the building combined with machine learning] to learn that people drink coffee in the morning so they should hire a barista."
Source: https://twitter.com/modestproposal1/status/1085583434788020227?lang=en https://twitter.com/modestproposal1/status/10855834347880202...
Sounds like a reasonable use of tens of millions of dollars of VC money.
- coredog64 7y agoNot a fan of WeWork, but there actually is something to this. We have two coffee makers in the break room at work, and there's always a line. If they brought in a barista who could work the machines, it's theoretically possible that you'd get better efficiency as there wouldn't be the dance of "who is next in line" or "sorry, gotta block the machine while I get sugar" or "I just want some hot water". I'd love to be able to point building management at some data that says that a barista costs $100/hour for 2 hours, but having them would unlock $1000/hour in lost productivity.
- roymurdock 7y agoTell your WeWork employed office manager (there is one in every office right?) to look at the line every morning when they're getting their coffee for two weeks. If the line is too long buy another 2 or 3 espresso machines, or supplement them with a few keurigs. I just saved you $20M per year in sensors, IoT platform subscription, and data scientist salaries for an investment of $5k. Pretty insane how much money is wasted on dumb BS just because there's so much money to waste and it's a trendy way to waste money Or more likely they wanted an innocuous reason to track all their office users at all times for other reasons
- ckorhonen 7y agoBut I’ve never actually seen a barista in any of the WeWork’s I’ve used (including the one which doubles as their HQ).