[nathancahill]

I'm curious about the shuttering timelines that failed startups use. Three weeks is awfully fast to go from operational to shutdown for a service that customers rely on. I see two options, either could be reasons that a startup would fail:

1. Customers don't rely on the service.

2. There are no customers.

Otherwise, why not leave the servers on for another month?

EDIT: 3. The service isn't automated enough to be run without human intervention, even for a month. Which would also likely create failure when it scales.

[chrissnell]

If your priority is returning the remaining investors' money to the investors, why would you keep the lights on?

[nathancahill]

I'm wondering specifically about keeping the servers on. Server costs are much less than the cost of "keeping the lights on".

[grrowl]

I agree, even shutting down compute/cost-intensive features (i.e. the actual functionality) and leaving the login screen open might prevent a few people from losing work.

Then again, if the code is mirrored on github/bitbucket/whatever, then the risk of data loss is minimal and the product is actually all functionality.

[javajosh]

It depends on the product. CI is very compute intensive (and may be bandwidth intensive, depending on setup).

[cognivore]

I keep hearing "fail fast" as part of the start-up rah-rah. They just took it to heart.