请问HN：你们进行A/B测试吗？
我在大型科技公司工作时接触到了极其简化的A/B测试。我想了解在大型科技公司之外进行A/B测试的情况。
例如，我最近与一家金融科技初创公司合作，他们并没有通过随机用户分组来进行真正的A/B测试，但他们确实有一个设置，可以在更大范围推出之前，先在特定用户群中测试某个功能。
那么，你们是否以某种方式进行A/B测试或功能标记？如果是的话，请你也提到以下几点：
* 每日活跃用户/每月活跃用户？我想了解在什么“规模”下A/B测试对组织来说是有意义的。
* 基础设施和架构是什么样的？是内部设置（在用户档案中添加标记或在代码中设置条件）还是支付使用某个SaaS服务来帮助？
* 维护成本如何？你觉得这样做值得吗？
如果你们没有使用A/B测试，原因是什么？是初期投资还是后续维护让人头疼？
查看原文
I was introduced to ridiculously simplified A/B testing while working at BigTech. I wanted to understand the landscape of running A/B tests outside of BigTech.
As an example, I recently worked with a Fintech startup that was not running true A/Bs via randomised user buckets, but they did have a setup to test out a feature with a particular cohort before a larger rollout.<p>So, do you do A/B testing or feature flagging in any way? If yes, can you also mention<p>* Daily Active Users / Monthly Active users? Trying to understand at what “scale” does A/B testing make sense for organisations<p>* What’s the infrastructure and plumbing look like? Inhouse setup (add a flag to a user profile or if conditions in code?) or do you pay for a SaaS that helps?<p>* What’s the overhead like? Do you feel it’s worth it?<p>If you are not using A/B testing, why not? Is it the initial investment or the followup maintenance that is the headache?