A Project managers perspective of Optimizely Feature Experimentation.
Use Cases and Deployment Scope
Our team uses it as a core part of our release and Validation process across client projects, not just internally. Currently, we are working with a whale client, a bank, and we've leveraged feature flags to test different credit card recommendation engines in prod without risking the live customer base. That has solved a longstanding problem of all-or-nothing releases, where any new algorithm introduced carried a huge rollback risk.
Pros
- We can do feature flag based rollouts with surgical control.
- It's good at progressive delivery in compliance-heavy sectors.
Cons
- We currently can't correlate feature flagged experiments with Salesforce CRM data.
Likelihood to Recommend
While Optimizely has genuinely elevated how we test and validate product ideas, it is without its pain points. The positives are clear, and I've alluded to them already, like clean experiment setups and reliable feature flags. But at the same time, scaling experiments across microservices and stitching data with external systems have proven to be a huge challenge.
