[jroseattle]

I remember when this article about the Etsy engineering department came out on Techcrunch 3 years ago.

https://news.ycombinator.com/item?id=9481377

The items that jumped out at me:

> The company owns and operates its hardware and networks in its own datacenter.

> The company has 685 employees of which approximately a third are engineers.

> It wanted to know how Hadoop worked, and the only way to do that was to bring it in-house and figure it out.

As a means to an end, this is a _really_ expensive way of operating nowadays. And when the business isn't rolling, these costs become magnified (and the associated operation vulnerable.)

[njyx]

I wouldn't necessarily agree with that. It depends on scale and the skills available to you. Yes, amazon and co do a good job of pushing down cloud costs, but you still need people to manage these services.

Also it depends WHEN you build things. Redshift for example is now pretty viable and potentially more cost effective than what Etsy does. 2-3 years ago that very likely wasn't the case.

I'm not saying Etsy doesn't have somewhat higher system costs than others but I doubt it's the cause of their issues.

[jroseattle]

> It depends on scale and the skills available to you.

Skills -- yes. Scale? That's kind of the point, nowadays. You don't need a large staff to accomplish scale.

> I doubt it's the cause of their issues.

Their eng/ops approach wasn't the cause of any business issues, AFAIK. That's why my comment is directed around what happens when business performance suffers -- you start looking for cost reductions.