Startups: scalability from the start?

4 points·kirbdee··5 comments

In implementing an idea/product would it be better to push quickly then iterate, or design and implement for scalability from the start?

Recommendations for scaling would help too.

Posted this on Quora as well: http://b.qr.ae/jWRWB7

View on HN

I think a middle ground is probably best. Choose a tech stack that can scale but doesn't slow down programming productivity much. Something like node, python twisted/tornado or Go-lang. Same idea for databases.

Using the right stack will usually get you up to a few thousand users without major headaches, after that you'll have to know what you're doing.

There are startups that fail to scale to a few thousand users, who are easy enough to find for a free service. It's not a nice problem to have when you're not making money solving it. A little scalability planning ahead helps, gives you more time to work on the product.

The general advice on HN is to ignore scalability at first. It's vague advice to a vague question.

If you can, you should always push quickly and iterate, looking for scalability later. That is fine for products/parts of products you can easily update or replace at low cost (e.g. software), but not for those you can't replace that easily and cheaply (e.g. hardware.)

So would something like creating the product on a simple lamp stack on an AWS instance be a good start?

Probably yes. Can it handle the first 1000 users? Then, just create and ship your product before worrying about scalability.

Indeed. It's better to discover if your product fits the market before trying to do the extra 80% of the hard work. If it fits, you have data to confirm, and you can go to a money raising round with solid information. If it doesn't fit, you can get out of something that doesn't matter quickly.

Reply on news.ycombinator.com