How we killed 80% of features and increased outcomes of a/b testing by 100% [eng]
Sometimes, less is more.
We often talk about how to design the system to be adaptable to changes, how to quickly add new functionality to the product. However, fewer people ask questions: “Do we need this functionality? Which features are really valuable for our end-user?”
In this talk, I’ll share our experience at Preply on how we approach the challenge of “Extracting the maximum business value, by keeping the product as simple as possible?”
We run more than 100 A/B tests simultaneously, meaning that there are no 2 users on the whole internet who would see the same version of Preply. I’ll share our vision on product development as well as technical details and challenges we faced during fostering and adopting the A/B testing culture at Preply. Among other topics I’ll cover:How to build an A/B testing system and product that will handle more than 100 A/B tests simultaneously?
- Can the A/B testing hurt the business?
- Hidden costs of running a/b tests
- Does it make sense to a/b test technical changes and refactorings?
- How A/B testing can protect you from losing several million $ on a simple library update?