We have built a great team to launch redesign mobile.
We are focused on big ideas that can change the world. We are also huge fans of great design.
In June, we had our first recruiting week.
Stay tuned for more.
In testing the new Amazon Local Register yesterday, my first transaction was for $300.
Amazon didn’t make any money on this transaction. (Nor will they make money on a lot of the transactions on Amazon Local Register at the 1.75% rate.)
There was a mistake in the original image of my transaction. (Sort of.) I needed to show the card and the original load value, so I made the card facing forward. In order for it to actually read the card, the mag stripe needs to be facing the camera.
The purpose of the transaction on Friday was to use up the remaining value of a PayPal debit card that I had purchased for test purposes. The PayPal debit card is a poor value compared with products like Walgreens Balance Financial and American Express’s Bluebird. T-Mobile customers also have a better value with T-Mobile’s prepaid card.
PayPal charges a monthly fee of $5. Rather than paying another fee, I decided to use up the remaining balance by charging myself with Square.
The total transaction amount was $3.75. Square did not make a profit on the transaction. (Come on, you know me! Did you really think I’d let Square make a profit?)
A number of people commented that I made a mistake by leaving my debit card number in plain view. I don’t make such mistakes. My credit card numbers in social media are like sideboob on TV: you think you’re seeing something, but you’re not. If you see a credit or debit card number, it’s a clue. (In this case, a clue that there was no money on the card.)
At first glance, this graph implies that U.S. cash transaction volume is plummeting. But look at the right side. The axis starts at $1,300 billion. If this graph had 0 on the Y axis, the line would essentially be flat. As posted, the graph is highly misleading.
Not using zero as the basis of the Y axis can be useful in certain cases, like analyzing short-term price movements in a stock. But this is a terrible use of it. Or maybe it’s a great use — because the writer wants to make a point unsupported by the data. But it’s still wrong.
Bonus error pointed out by a reader: The “We Are Here” line shows us between 2013 and 2014. We’re between 2014 and 2015.
I received the catalogs from Crate and Barrel and CB2 on the same day.
It’s better to spread that delivery out over multiple days because it increases the likelihood that I’d look at one of them. On the same day, it’s possible that I’m on vacation and come back to a pile of mail, my (hypothetical) spouse picks up the mail and tosses it out, etc. Spreading the delivery out doubles the chance that I’ll look at it.
To address some other comments from readers:
I was waiting for an Uber and the driver zoomed past me at 35 miles an hour in downtown SF. (Why you wouldn’t drive slower when you’re approaching a passenger is beyond me.) When I flagged him down on the one way street, he backed up in heavy traffic to get to me.
That is the kind of driver that shouldn’t be on Uber’s platform. When I reported what happened, Uber took note and said they were going to reach out to the driver.
But the current process requires active input from passengers. Unless a drive is egregiously bad, most people wouldn’t bother.
Technology provides an easy answer to the problem: passively tracking driver behavior. If someone has a lot of quick stops, swerves a lot, brakes hard or speeds down city streets, that person shouldn’t be on the Uber system.
There’s also another huge advantage: you can track driver behavior when they don’t have a passenger but are logged on the system. The current Uber model provides a strong incentive for drivers to be reckless when they don’t have a passenger.
For the level of detail that Uber needs to bump bad drivers off the system, the sensors on modern phones are great.
There are technologies that provide even better data. A start up called Automatic sells a device that consumers can plug into the OBD-II port on 1996 or later cars. The app warns you when you are speeding and braking hard. it also provides logs of your driving and average miles per gallon.
My auto insurance is from Metromile, which uses data from the OBD-II port to charge me by the mile instead of typical pricing models. (I estimate that I’ll save 30% off my former GEICO rates.) Incidentally, Metromile won’t provide coverage while you’re providing taxi services with Uber.
One survey respondent suggested deactivating the Uber app while the vehicle is in motion, much like some embedded navigation systems. Unfortunately, the app is so dependent on people interacting while driving that that won’t happen.
Many taxi drivers are reckless, too. But we can’t do much about that.