# When to ship it, when to kill it

DevFeed: [When to ship it, when to kill it](<https://devfeed.tech/articles/when-to-ship-it-when-to-kill-it-41172.md>)

Original publisher: [Read original article](<https://www.craigkerstiens.com/2014/08/13/When-to-ship-it-when-to-kill-it/>)

Author: Map

Published: 2014-08-13T20:55:56Z

Content type: opinion

Language: en

Sources: [Craig Kerstiens](<https://devfeed.tech/sources/craig-kerstiens.md>)

Topics: [Heroku](<https://devfeed.tech/topics/heroku.md>), [Users](<https://devfeed.tech/topics/users.md>), [Testing](<https://devfeed.tech/topics/testing.md>)

Tags: [beta](<https://devfeed.tech/tags/beta.md>), [feature](<https://devfeed.tech/tags/feature.md>), [flag](<https://devfeed.tech/tags/flag.md>), [heroku](<https://devfeed.tech/tags/heroku.md>), [philosophies](<https://devfeed.tech/tags/philosophies.md>), [product](<https://devfeed.tech/tags/product.md>), [testing](<https://devfeed.tech/tags/testing.md>)

## AI overview

The article presents a product-shipping framework based on alpha or beta testing with users. It recommends gradually rolling out a feature, temporarily removing it or disabling its feature flag, and using immediate user feedback to decide whether to improve or kill the feature.

## Source excerpt

A few weeks ago at lunch I had the opportunity to catch up with a company in the current YC batch, building something very similar to dataclips. While we talked about a lot of things from what we've learned from dataclips, marketing, and other areas. One area we talked about was product and when to ship vs. when to kill things and I realized I hadn't talked on my fairly simple but clear view on this publicly, so here it is. A large credit to Adam Wiggins for giving this model early on in Heroku and his approach to shipping product. A precursor to shipping First a little background on shipping, in shipping something I'm going to assume you have some process of alpha/beta testing with users. This is actually fairly key, if you're not testing it with users then well the rest of this is all moot. Alpha and beta testing is pretty simple, you need some early users. These can be friends, people within a network, or random users you select from. There's different value to how you select these but that's a topic for another time and place. On to shipping So how do you know it's ready. The basic idea is super simple. Give it to some users in alpha/beta testing. Or start to roll it out following a one -> some -> many all principle (maybe to 5% or 10% of your userbase). Then take that brand new feature away. There's a couple of ways to do this as far as mechanics. If you're in contact with users such as alpha/beta users that you were higher touch with just email them. Tell them you're removing the feature, or if you want to approach it more softly ask them how much they'd miss it if it were gone tomorrow. If you're rolling it out more broadly perhaps behind a feature flag, flip it off and watch for feedback. Once you take the feature away or threaten to if you don't have users with pitchforks almost immediately then it's not ready to ship. Go back to the drawing board and work more on it or simply kill it. As @james_heroku would say: "So you're saying the reason to ship the shi