Graham McNicoll

Engineer's who ship products without measuring them are making a big mistake. If you aren't running controlled experiments, you have no way to prove if a 15% latency spike or a churn increase was caused by your latest updates. There are so many over variables that happen that can effect metrics you need to use controlled experiments to determine causal impacts. Without scientific rigor, your team is just "observing" noise. Running controlled experiments (A/B tests) are the best way to determine impact. Without experiments, you'll end up with an unexpected change, and then sorting through a mountain of PRs trying to find what caused the change. I built GrowthBook to make it stupidly easy to measure everything you ship.

Only one snapshot so far — the engagement-over-time curve appears once the daily scrape has captured this post at least twice.