PostHog made huge strides in 2021. To name just a few landmarks, we:
All of this (and more) lead to huge growth. Users grew by 4x, the number of open source contributors doubled and we acquired numerous reference customers.
But this is just the beginning. We've also been busy making PostHog a broader and more powerful tool for making great products.
To celebrate, we're officially launching a new feature or initiative every day this week. Welcome to PostHog Launch Week I: A Universe of New Features.
As the biggest dogfooder of our own product, we were beginning to see the consequences of the lack of in-house tools to manage our data. All the use cases we saw boiled down to a few common pain points:
We set out to solve these pain points and the result is the new Data Management experience we're introducing today on PostHog Cloud and next week for self-hosted.
Read: Introducing Data Management for PostHog
We created 'Project Sparkle' to address two challenges: to make PostHog scalable for anyone and to make self-hosting PostHog easier.
In The secrets of PostHog query performance, Karl-Aksel Puulmann details the work we've done to make PostHog faster and more scalable, including (among other things) a massive 55% improvement to query performance on PostHog Cloud.
In How we’re making PostHog deployments easier, Harry Waye and Guido Iaquinti explain how PostHog's architecture has evolved since its launch in 2020, and the steps we're taking to make self-hosted deployments easier, including multi-layered testing and improved monitoring.
Finally, we're delighted to officially confirm our partnerhship with Altinity and the launch of PostHog Marketplace. You may know Altinity as experts in all things ClickHouse and data infrastructure. Altinity will be offering a variety of services to help give large organizations on PostHog an increased level of support for self-hosted deployments.
Going forward, the marketplace is where you’ll be able to find third-party services which layer on top of PostHog, such as companies offering support, integration or development services. Altinity is the first such company, but there are more to come in the future.
Everyone has an opinion about user experience, and that's fine. But it's vital to test those opinions and validate what works, which is why we built our Experimentation suite.
Experiments allow you to choose a target metric, choose specific people to run this experiment on, and set how long the experiment runs for.
Thanks to Feature Flags, you can then easily validate whether each variant looks good, launch your experiment, and wait for data to come in. We run a Bayesian analysis on the data to give a probability for each variant being the best, show a graph of how things are looking for each variant, and whether the results are statistically significant or not. It's awesome.
Today, Neil Kakkar shares three things we've learned about running effective A/B tests. If you want to build better products, we seriously recommend giving it a read.
Read: What launching Experimentation taught us about running effective A/B tests
95% fresh on Tasty Earthworms
"So good, I wrote a 1,000 word Twitter thread about it" - A Twitter user, probably
Enjoyed this? Subscribe to our newsletter to hear more from us twice a month!
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。