Run Experiments Against New Features With DevCycle

What happens when you have a feature that you’re developing behind a feature flag and you want to run some targeted experiments? If you’re using a traditional feature flagging system you may be out of luck, or you may need to transition to a dedicated A/B testing tool such as Taplytics.

Today we’re launching Experiments for DevCycle

Built right into DevCycle’s market-leading audience targeting tools is now the ability to transform any feature flag into an experiment. All you need to do is set a targeting rule to serve a Random Distribution!

That’s all you need to do and you’ve created an experiment. The experiment could be targeting all users, or some group of early adopters you’ve identified in your user base. Either way, DevCycle takes care of ensuring distribution is properly randomized so you can get great insights from the data.

Why Experiment?

Now given we’re a developer tool there may be questions about why to experiment at all. Most teams just want to use feature flags to manage risk and gradually roll out new features while driving efficiency and quality up.

That’s ok, but at some point you or your team may want to expand the reach of how you use feature flags into the realm of experimentation for a variety of reasons.

You may want to experiment on any of these things and more:

• Validate to make sure application performance remains the same or improves.
• Validate in a controlled way whether code changes increase or decrease error rates.
• Confirm that a new feature is actually driving more conversions or revenue.
• Measure real impacts of features to SLAs and SLOs.

All of these and more are great reasons to start thinking of feature flagging and experimentation as part of a virtuous cycle of building, shipping, testing, and iterating.

Start Small

If you’re ready to start running some experiments against your new features sign up for free today! And check out our docs on Randomized Distribution.