I'd be really interested in the real-life lifespan of scooters, including theft, ditching (cue pictures of piles of ditched e-bikes sitting in a stream), customers doing stupid things and breaking them etc. Also on real-life utilisation percentage (ride length, rides/day). Together these could make an order of magnitude change in the profit expected.

The article is making the point that changing the key parameters dramatically varies your expected profit, but the message I'm taking away is you can make any business model look arbitrarily good or bad by slightly tweaking your estimates, and the more multipliers in your model, the more 'flexible' it is in this way.

I think it's great that the tech/startup industry as a whole has moved more towards a "let's ensure the unit economics work" mentality vs. the "we'll grow quickly and monetize through advertising later" mentality of ~5-7 years ago.

"Getting lots of users" is not a viable business model.

This is a nice analysis. TechCrunch has been doing much better work recently.

Always wondered, given the profitability, as shown in this article (mainly due to Chinese electric scooters being cheap) - why no scooter sharing startups in London?

Does the subsidised bike sharing scheme make anything else unattractive?

I’m sure it’s coming... bird just launched 7 months ago and raised serious money in the last month, and all the copycats just launched last month!

I know this is a deliberately simple model, but treating "custom code" as a variable cost that increases with each unit sold makes me sad.

I thought that per-unit cost was for the GPS/3G hardware plus whatever custom circuit board (or Arduino or whatever, heh) they use to tie it all together. The "plus some custom code" bit was just a nod in the direction that some software would be required, even if it's not been assigned a number.