Building and shipping exceptional product experiences propel businesses forward. They drive higher customer satisfaction and loyalty, increase conversion rates and customer lifetime value, and improve overall market share. That’s why companies invest in building and shipping great features in the first place.
But how do you know the features you’re shipping are producing the desired outcomes? Well—you don’t. Unless you are measuring the impact of every feature that’s released and, also, making sure it’s providing optimal value towards business goals.Â
The good news? With LaunchDarkly, you can run experiments on every release to help ensure, with data and analytics, that the features you shipped had a positive impact on your business. Software teams (engineers, product owners, and data scientists) can work together to embed experimentation into the software release process, enabling them to better measure and optimize features at scale, deliver winning product experiences faster, and ultimately drive progress against key business initiatives.Â
The even better news? We’ve released some pretty neat features to better help experimentation teams measure the impact of every feature, conduct statistically rigorous (and flexible) experiments, and deliver winning experiences to customers faster.
Measure the impact of every feature
We know it’s important to not only control how features are being released but also to measure the impact features have against key business objectives. That’s why we increased the power and flexibility of our metrics you can set up in LaunchDarkly so that you can answer (and track) the questions and metrics that matter most to your business. We are also making metric verification easier than ever for your team to help confirm your implementation is working as expected and your metrics are tracking as desired. Â
Flexible metric design
Precisely configure each metric with a guided setup that allows you to define exactly how LaunchDarkly will accept and aggregate your measurements to answer the exact questions your team needs answers to.
Metric setup verification
Easily verify your metric statuses in LaunchDarkly with simple visual cues and detailed lists of received events so your team can confidently run your experiments. Â
Conduct statistically rigorous (and flexible) experiments
We also know it’s important to run the types of experiments that matter most to your business. That’s why we are adding additional statistical methodologies into LaunchDarkly Experimentation so that you can confidently run sophisticated experiments that help deliver trustworthy outcomes.
Experiment holdout*
Easily compare the long-term lift on your metrics across many experiments with a control group that is deliberately excluded from experiments.Â
Mutually exclusive experiments*
Leverage experiment layers to control experiment audiences and ensure experiments running in parallel do not have overlapping participants.
Deliver winning experiences to customers faster
Great—you are measuring feature impact and running cool experiments, but what really matters most is shipping great product experiences to your customers as quickly as possible. That’s why, with LaunchDarkly, we are making it easy to reach experimentation results faster, visualize experiment impact against your key business metrics, and determine when the results of an experiment are stabilized so that you can confidently ship a winning experience.
CUPED*
Reach trustworthy experiment results faster with less data by leveraging CUPED (Controlled-experiment Using Pre-Experiment Data) right within LaunchDarkly.
Credible interval time series
Visualize how an experiment’s results are stabilizing over time to better understand when you can confidently make a decision and ship a winning treatment as early as possible.
Relative metric lift summaries
Quickly understand how different experiments are impacting key metrics so that you can make informed decisions that best support your business objectives.
But wait, there’s AI tooÂ
Enhance experimentation setup and design with LaunchDarkly’s AI-powered experiment builder.Â
Now, when designing an experiment within LaunchDarkly, you can leverage AI to streamline setup and suggest metric and flags variations to ultimately optimize experiment design and produce winning experiences for customers. Customers wanting to leverage LaunchDarkly’s AI-powered experiment builder can reach out to their account teams to request access.Â
Measure and optimize your features nowÂ
Ready to take the leap to embed experimentation right into your feature release workflow with LaunchDarkly?Â
To get started experimenting today (or simply learn more) request a demo today or check out our product documentation.Â
--
*features coming soon
