Case Study: Duolingo

Duolingo has a huge number of users and made the decision to start working with an MMP to maximize the impact of their marketing and make stronger, data-driven decisions. After undertaking extensive research and engaging in peer reviews, the engineering team settled on Adjust.

Before working with Adjust, the team at Duolingo had no clear UA strategy. An engineer would “throw up”simple campaigns and Facebook ads and track their performance at a basic level. When Arthur Jun, Head of Performance Marketing & Media joined in 2018, he was also tasked with setting up a more sophisticated performance marketing team and establishing the infrastructure and tech stack. Ensuring event data was accurate, that optimization events for partners were integrated properly and to continually test, plan and strategize for future marketing campaigns and activities was integral to his approach.

Duolingo is, first and foremost, an engineering and product-driven platform, so the goal was to workout how to optimize it for marketing performance by seamlessly integrating a tool from an engineering perspective. In essence, they asked themselves how they could use Adjust to take their UA campaigns to the next level and help drive their overall success.Moving forward, Arthur and his team are looking to use Adjust to further understand their user lifecycles- to better optimize their campaigns, driving Duolingo Plus subscriptions and tie brand awareness in with performance marketing metrics.

With the key issue being performance measurement of UA campaigns and understanding user journeys, choosing Adjust as their MMP and single source of truth has empowered the Duolingo team to make data-driven decisions that drive campaign performance forward. TheAdjust CSM team also provided one-on-one consulting solutions to help train their engineers in understanding how the product works, allowing them to effectively report, analyze and accurately measure performance.

By using Adjust’s measurement and analytics tools for optimization and analytics, they’ve significantly improved their overall efficiency.With their datasets now in one place, they’re able to make holistic, informed decisions. With the level of data visibility and granularity provided, the team has been able to focus on the KPIs that matter to them most - day 7 retention, monetization events, one-year LTV, and three-year LTV.

Working with the Adjust Fraud Prevention, which the team are currently testing, is already allowing them to proactively fight ad fraud in real time. It’s also helped them to define what programmatic means for Duolingo, and to have a stronger framework for understanding the types of fraud that exist, where it comes from and how it impacts and skews data sets

Duolingo’s journey with Adjust enabled a restrategizing of how they work with user acquisition and measurement, which packaged together meant aswathe of actionable insights and measurable results.Using Measure, their day 7 retention and LTV improved drastically, as they were able to understand the metrics at the acquisition level - without Adjust, there was no attribution to an actual media source

As Duolingo’s single source of truth, Adjust empowered the team to become more agile by:

• Improving their relationships with partners including Google and Facebook. Now they better understand the metrics they should optimize for and how to make their general performance portfolio better.
• Helping them to gain a common ground for a company-wide approach to marketing.
• Enabling them to integrate ‘awareness media’ campaigns using Adjust tracking URLs when launching, for example, newsletters with a partner/affiliate, or large branding campaigns.
• Providing them with deeper insights into channels that are working vs. not working as they trial a lot of new acquisition channels - including influencer marketing.
• Proactively stopping ad fraud, allowing them to rely on clean datasets and see which channels are actually performing best.

Duolingo started at Carnegie Mellon University in Pittsburg has a project by a university professor, Luis von Ahn, and his graduate student, Severin Hacker. It was created with the goal of teaching language for free through technology.By 2011, the team had raised USD $3.3 million, launched a private beta, and gathered more than 300,000 waiting list users. The app launched globally to the general public in June 2012, and as of 2020 has a community of over 500 million total users and 40 million monthly active users based in nearly every country in the world.

The app, which lies at the crossroads of the education and gaming verticals, features a reward system and a progress tree. There are also leaderboards where users can compete in their chosen languages and gain badges to display their progress and levels. The process of acquiring languages and studying in-app combines methods of listening, speaking,reading, and writing in short game-like exercises.

In addition to its millions of at-home users, it’s also a favorite in schools and businesses. Most of its features are free to use, with the opportunity to remove in-app advertising by upgrading to Duolingo Plus.

• English is currently Duolingo’s most studied language, followed by Spanish, French and German.
• Japanese and Korean are the top two Asian languages.
• In Ireland and Sweden, their own languages are the most studied on the platform.
• Spanish is the most studied language in the U.S., followed by English.
• Navajo and Hawaiian also attract a lot of learners, especially for brain training and culture, and Hawaiian even reaches the #2 spot in Hawai'i.