October 2024 Summaries

Retail and eCommerce companies face significant risks from software failures, which can lead to lost revenue and damaged brand trust. To mitigate these risks, five key strategies can be employed: decoupling deployment from release using feature flags; progressive delivery for incremental feature rollouts; kill switches for immediate rollbacks; dynamic configuration for flexibility in changing conditions; and automated monitoring and remediation for proactive issue resolution. These strategies enable companies to balance innovation with stability, ensuring smooth software rollouts and maintaining customer satisfaction during high-demand periods.

The blog series discusses how GenAI is transforming software delivery and introduces new challenges such as non-deterministic outputs and a rapidly shifting model landscape. It highlights the differences between GenAI and traditional software development lifecycle (SDLC) in planning, design, implementation, testing, deploy, and maintain phases. The article suggests that feature management can help teams build and deliver new GenAI features within hours, not weeks. LaunchDarkly is working on supporting GenAI builders with upcoming features to support the new GenAI software development lifecycle.

LaunchDarkly has introduced enriched experiment analytics, providing users with deeper insights into the results of their feature experiments. The new features include total conversion and exposure columns for a complete picture of user behavior and feature performance, statistical calculations to validate results, and expected loss calculations to minimize risk. These enhancements enable faster decision-making, increased confidence in results, improved report generation, and reduced experimentation risk. Enriched experimentation analytics is now available to all LaunchDarkly Experimentation customers.

This tutorial demonstrates how to run an A/B experiment using LaunchDarkly and ExpressJS to test the impact of seasonal CSS on signup conversions for a costume shop website. The process involves creating a metric, setting up an experiment with flags, adding event tracking and flag evaluation to the ExpressJS routing layer, and observing the results in the LaunchDarkly app. This method allows developers to gather data about the impact of changes to their product and make informed decisions based on that data.

DORA (DevOps Research and Assessment) Metrics are four key performance indicators used to evaluate the efficiency of software development teams. These metrics include deployment frequency, lead time for changes, change failure rate, and mean time to restore. By measuring these metrics, organizations can gain insights into their DevOps performance, identify areas for improvement, and make data-driven decisions to optimize their software delivery processes. Implementing DORA Metrics requires commitment, collaboration, and a shift in mindset, but it ultimately leads to faster feature releases, better user experiences, and more efficient development teams.

The text describes a tutorial for creating a dynamic "Link in Bio" application using Reflex.dev and LaunchDarkly. The author explains how to set up the tech stack, create a customizable link in the bio application, and add feature flags to switch between different versions of the page. The tutorial also demonstrates how to integrate LaunchDarkly SDK into the application for managing feature flags. This versatile "Link in Bio" app can be used by anyone showcasing their projects or online presence, allowing them to easily adapt to various audiences and scenarios.

This tutorial demonstrates how to build an ExpressJS demo app that combines Twilio Verify, a fraud verification API, with LaunchDarkly feature flags to toggle extra verification layers on during spam attacks. The app allows users to sign up and includes a signup flow where the user enters their phone number during signup and receives a verification code via SMS. A LaunchDarkly flag is added to conditionally enable SMS verification, allowing developers to adjust the signup process based on local conditions such as spam spikes.

Experimentation plays a crucial role in attributing KPI movements to business value. It helps create robust data about user reactions to new features and allows for the identification of incremental value or risk avoided. By randomly assigning users into control and treatment groups, businesses can ensure that results are not influenced by external factors. Experimentation also enables quantification of experiment value, which can be used to calculate ROI and prioritize feature development based on customer needs. Even small experiments contribute to overall business growth as they provide valuable learning opportunities and help in making data-driven decisions.

This four-part blog series explores how Generative Artificial Intelligence (GenAI) is transforming software delivery and the challenges it introduces. GenAI is changing the types of software being built, with many product teams now asked to "build with AI" in their core products. However, taking non-deterministic applications to production at enterprise scale is complicated and risky due to factors such as rapid innovation in LLMs, pressure to improve models, and challenges in finding optimal prompt & model configurations. Balancing safety and reliability while constantly optimizing for state-of-the-art experiences is critical as GenAI features become expected in modern software experiences. The next blog will cover how GenAI development differs from the traditional SDLC and the types of challenges and opportunities it introduces.

This tutorial demonstrates how to use LaunchDarkly's progressive rollouts in an ExpressJS application for upgrading to a new API version. It covers setting up the required environment and dependencies, creating a flag in LaunchDarkly, configuring a progressive rollout, and monitoring the progress of the rollout. The example used is an API for raccoon facts, with the goal being to switch from Version 0 to Version 1 gradually while monitoring for any potential issues. This method allows developers to control the release pace and mitigate risks associated with new features or updates.

In this tutorial, developers learn how to reduce risk in their AI-powered apps by implementing rate limiting with Arcjet and leveraging LaunchDarkly's targeting capabilities. The text guides users through setting up the project using Next.js, Arcjet, Replicate, and LaunchDarkly. It also covers adding rate limiting with ArcJet to save money on AI bills and configuring LaunchDarkly for proper context from LaunchDarkly.

Bugs in software can cause disruptions, but Progressive Rollouts can help mitigate these risks by incrementally exposing new features to segments of a user base. This phased approach allows for gradual increases in traffic exposure over time, limiting the number of users affected by any given issue with the release. Progressive Rollouts are now an option on a flag's targeting rule in a given environment and can be set up with various percentage and time increments to control user exposure. This feature is available for all LaunchDarkly plans and can help safeguard release processes and ensure that issues only affect a subset of users at most.

In this guide, we'll build a dynamic email marketing system that personalizes emails for different user segments using Resend for sending emails, LaunchDarkly for controlling feature flags, and SQLite for storing user data. We'll learn how to use LaunchDarkly to manage feature flags and personalize email content, send customized emails at scale using Resend, pull user data from a database, and tweak our email strategy in real-time with LaunchDarkly's flexible rules. The entire project code is available on GitHub.

An AI-powered application designed to generate wilderness survival scenarios for a reality TV show, inspired by "Naked and Afraid," is developed using Next.js, LaunchDarkly, and Replicate. This app lets users create amusing scenarios by selecting celebrities and environments, utilizing the Replicate API to generate content. LaunchDarkly enhances the app's functionality by allowing seamless testing of different prompts through JSON feature flags without redeploying the application. The app demonstrates scalability and flexibility by enabling prompt and model adjustments in real-time, creating entertaining scenarios like Nicolas Cage surviving in a giant burrito. The tutorial concludes with a promise of further enhancements, such as rate-limiting API requests, in an upcoming continuation.

LaunchDarkly's free plan now includes progressive rollouts, allowing developers to release features gradually and monitor for any issues that may arise. This tutorial demonstrates how to use this feature in a Python FastAPI application to upgrade to a new API version. By gating updates with a LaunchDarkly flag, developers can control the pace of their releases and ensure a smooth transition to the latest version. The example provided showcases how to set up a progressive rollout for an API update, allowing users to see new features at a controlled rate while minimizing potential risks.

Financial services companies can tailor digital financial services through continuous experimentation, leading to increased customer engagement and improved business outcomes. By leveraging tools like LaunchDarkly for controlled tests and A/B testing, institutions can gather data to fuel personalized experiences while maximizing efficiency and reducing guesswork. Experimentation helps teams focus on features that customers want, reduces wasted effort, and improves overall product development efficiency.