Ritual delivers customer experiences with experimentation

Ritual is a personal health technology brand building the future of daily essentials. The brand was founded on the belief that better health begins with better ingredients, and has pioneered a new standard of high-quality, clean products that are backed by science and Made Traceable™ with the first visible supply chain of its kind. Together with some of the world’s leading scientists, researchers, and advisors, Ritual has developed products based on thousands of independent research studies, has completed a peer-reviewed clinical trial on its flagship multivitamin, Essential for Women 18+, which also received USP Verification and Non-GMO Project Verification. Ritual products are trusted by hundreds of thousands of customers, and its multivitamins and protein are formulated to help support foundational health for many members of the family including men, women, teens, and kids. Learn more at Ritual.com.

Ritual's digital product delivery team operates at the cutting edge. They built their own state-of-the-art tech stack. They eagerly adopt modern development practices. And they strive to be data-driven in everything they do.

True to this ethos, Ritual’s developers wanted to make feature management a core piece of software delivery for their e-commerce website. Not only that, they wanted to weave experiments into this same workflow.

At the time, product managers at Ritual were already using an experimentation software provider. Aware that this tool included free feature flags as a part of its overall package, the development team chose to give it a try.

Developers were disappointed with the lack of feature management capabilities in the solution. And they found it problematic that feature flags and experiments required separate implementations.

Every time they wanted to experiment on a feature, it led to a morass of Jira tickets. They already had to do the work of building the feature itself. But then they had to separately implement an experiment and remove it once complete. What’s more, upon identifying winning variations from experiments, developers had to rebuild the features in code. This was painful.

The disconnect between flags and experiments slowed the team down, resulted in fewer experiments, and made it tough for developers and product managers to stay aligned.

Daniel Archer, VP of Engineering at Ritual, explained:

The development team thought about constructing a feature management and experimentation solution in-house. After all, they’d built their own tech stack. But they reasoned that doing so would detract from their main business.

Ritual chose LaunchDarkly because they saw it as the most robust, developer-friendly platform on the market. And it supported experiments and feature flags in a single implementation.

The team was extremely pleased with how LaunchDarkly entwined flags and experiments.

It’s now far easier for the company to start, launch, and clean up experiments. Moreover, developers no longer have to rush to scrub experiments from their code.

Before, with the old tool, if they wanted to release a winning feature to a broader audience, they had to leave the experiment associated with that feature in the code for routing purposes. The longer the experiment remained, the more impressions it racked up. Unfortunately, Ritual got charged for these impressions. Thus, any time developers needed to roll out a winning feature, they first had to race to clean up the experiment and then redeploy the feature.

Whereas, with LaunchDarkly, once they identify a winner, they can just disable the experiment and finish the release without incurring extra fees.

Ritual has greatly increased the efficiency of implementing and launching experiments. They've shortened the time it takes to move from one experiment to another. And they run such tests in production safely.

Now that feature flags and experiments are baked into the development workflow, engineers can deploy winners to 100% of users without making any code changes. They just make a simple flag targeting update in LaunchDarkly. In some cases, they use Scheduling (a part of Feature Workflows) to release automatically.

Ritual employs flags and experiments for a host of critical use cases like conversion and retention optimization. They also leverage fine-grained targeting. Developers and product managers roll out features to discrete user segments based on browsers, devices, and geo-attributes.

Notably, some of the organization’s most avid LaunchDarkly users hail from outside the software engineering team.

Product managers, quality assurance (QA) engineers, and other business partners use LaunchDarkly to test new features internally, run sophisticated experiments, and dark launch features with little developer involvement. They often run multiple experiments in parallel.

As an example, they created two new versions of an upsell feature and conducted experiments to measure product add rate for each. With LaunchDarkly, they also adjusted the rollout percentage for each variation. After analyzing the results, they released the winner to all site visitors. It was effortless. And it made a direct impact on the bottom line.

The team also uses LaunchDarkly's Data Export Add-on and native integration with Segment to pump feature flag event data to their business intelligence (BI) tools. They collect, collate, and analyze LaunchDarkly data alongside the rest of their marketing data.

LaunchDarkly empowers all teams, starting with developers. Ritual’s experience affirms this notion. Their developers were delighted with the many capabilities in LaunchDarkly geared toward software engineers. Developer satisfaction and productivity have risen.

They now ship code to production frequently, safely, and confidently. And the wider product delivery team releases that code to end-users whenever they please.

Before, Ritual deployed to production roughly twice a week. Now? Dozens of times a day.

LaunchDarkly enables Ritual to devote their best energy to shaping the future of health technology. Collaboration between developers and product managers has never been better. And the company is more data-driven, which has had an appreciable impact on revenue and the customer experience.

When it comes to their elite approach to software delivery, Ritual breathes rarefied air. LaunchDarkly is grateful to play a part in that.