Statsig

It's Experimentation Week - Day 2 ✌️ Today, we are announcing one-click Stratified Sampling to make your experiment results more trustworthy and consistent.  Data Scientist Craig Sexauer explains how this feature works in Statsig and helps you reduce false positive rates at the same power. Shoutout to Nick Jiang Minhye Kim Timothy Chan who also worked on this! - Stratified sampling is especially useful for experiments where a tail-end of power users drive a large portion of a metric value, such as in B2B experiments with small sample sizes. - All you do is provide a metric to balance on (like contract size or number of employees) or a classification (like subscription tier or industry sector) during experiment setup, and we'll balance the control and test groups.  Note: This feature is available on both Statsig Warehouse Native and Statsig Cloud. That's 2 down, and 3 more to go! Keep following us throughout the week to see what else we're shipping 🚀 #experimentation #analytics #datascience

Statsig continues to grow, and we're looking to hire our first Growth Data Scientist. This individual will lead the analytics effort on our product-led growth (PLG) initiative by working cross-functionally across Data, Business Operations, and Marketing. You will find opportunities to improve user acquisition, onboarding success, and provide actionable insights for our marketing and product teams to assist our PLG efforts. This role is 100% in-person in Bellevue, WA. Interested? Application link is in the comments.

Experimentation Week is here! 📌 Day 1: We’re kicking off the week by announcing Meta-analysis. Head of Data, Timothy Chan, explains that many customers have been using Statsig long-enough to have accumulated a solid corpus of experimental results from which to derive valuable insights. We’ve shipped two powerful views that enable meta-analysis: 1. Experiment timeline view: Analyze insights from all your experiments, filter down to a team’s experiments, and instantly get answers to questions like: What experiments are running now? What percentage of experiments ship control vs. test? How much do they impact key metrics? 2. Metrics correlation view: Visualize and scan for correlations across metrics. Quickly plot two metrics on the same chart to check whether they tend to move together. We’d love for you to try this out and hear your feedback! We have more coming here too. Some examples: Near term, we’ll add a feature to look at “batting averages” for metrics you’re trying to move. Mid-term we want to make it easy to back test new metrics against old experiments to validate sensitivity. We’ll be announcing a new feature each day this week. Follow along on this feed and stay tuned 👀 #experimentation #analytics #datascience

Three months ago, I decided to join Statsig as the first customer data scientist. Many people have asked me how I made this decision. It comes down to three things: 🌻 I know the job is hard but valuable. Data is a powerful weapon in today’s world, from developing new medicines, to evaluating AI models, or driving viral product growth. However, data is often underutilized, relegated to being some ‘fancy’ charts in executive reports. I believe in Statsig’s mission; once we help people unlock the value of data, we can drive unprecedented innovation across different industries. 🏆 I am proud to be part of this team. I remember how eye-opening it was to attend my first standup at Statsig - the company has daily standup where everyone shares their works and get applauds. Since then, every day I have been impressed with how many cool things are shipped at Statsig. One great example is Statsig Warehouse Native (WHN), the most popular WHN experimentation product today - https://lnkd.in/dBX7FxpZ 🚀 Just hop on the rocket ship. A new job should be thoroughly evaluated from many dimensions, but at the end of day, it still takes a leap of faith. Like the famous Silicon Valley quote: “If you’re offered a seat on a rocket ship, don’t ask what seat. Just get on.” Now I am onboard. I am super excited for the wonderful adventure ahead! If you want to chat about experimentation, data science, startups or something else, feel free to reach out and lets grab a coffee or bubble tea 😊

Do you love Statsig as much as we love Statsig? Now you can rep it... From stickers to hoodies, we've got you covered. PS. I just ordered the pixelated Statsig dog stickers - Ross Geller will be featured on my laptop now! visit: www.statsig.store Big thanks to Cat Lee and Jessie Ong, the Statsig Brand Designers! #statsig #statsigstore

👀 Less than 3 days to go until Experimentation Week! 🌟 Follow this page, and starting Monday, we’ll be announcing a feature a day throughout next week, showcasing some of the advanced experimentation and statistical methodologies the team has been working on. Five days. Five launches. Get ready 🚀 #experimentation #analytics #datascience

Happy Feature Friday! Loads of new features are shipping this summer 🌞 😎 📈 Week 23: Check out our new Formula mode that makes your Product Analytics explorations easier. - Now, when you add a formula to a chart, all other series will be hidden by default. This is especially useful when adding series to create percentages, as these values are often much smaller than the input series. - You can also save charts or add them to dashboards with specific series hidden. This makes it easier to view and share the results of your formula-based explorations, keeping your visualizations focused on the most relevant insights. 🤩 Try the new formula experience today and stay up to date with everything we’ve been shipping over the past few weeks across our product lines (feature flags, experimentation, analytics) here: https://lnkd.in/g4D7Q9N3 #analytics #experimentation #dashboards #startups

One little-known (and rarely discussed) problem in experimentation and feature flagging is the impact of ad and content blockers. These can end up blocking feature flags, experiment data, and even runtime dynamic configuration, resulting in bugs or the exclusion of users. This can impact test results by reducing statistical power or potentially introducing bias. We just rolled out a new feature to solve this problem: a managed API proxy. This gives you a unique URL to send and receive data to/from Statsig servers, which makes it way less likely to be intercepted by client-side or DNS blockers. We've always supported and documented how to set up your own proxy, but this one click solution is super easy to set up and available to all pro tier customers. Full docs page link in the comments! Thanks Matt Garnes for bringing this magic to the Statsig console!

Phone booths aren’t just for people, they’re also for barking dogs when they need a timeout… Elizabeth George’s Ross Geller goes straight to jail 🐶 #statsigsnapshots

Statsig ❤️ Golf ⛳ Our Startup of the Week, Live Tourney, provides easy-to-use golf tournament and live scoring software for golf courses and tournament organizers. Their Founding Engineer shares how Statsig helps them decouple code changes from feature rollouts: "As our customer list grows, it's important that as the only Engineer, I don't have to make code changes to roll out new features or turn off something that may have an issue. We roll out new features to a happy set of customers first and find issues before rolling them out to our broader customer base." Before Statsig: "I'd just ship any changes to everyone and cross my fingers that I did all the right testing, but inevitably, some edge case would come up and I'd have to spend my weekend urgently fixing the issue." After Statsig: "With Statsig and a recent critical change, I found two issues that would have been quite bad before the majority of our customers even had the change. I'm looking forward to getting my weekends back." Join developers and startups like Live Tourney in building awesome products with Statsig: https://lnkd.in/gBiWvKuj #startups #developers #featureflags