Behind the product: Duolingo streaks | Jackson Shuttleworth (Group PM, Retention Team)

Jackson’s background and an overview of Duolingo’s streak feature (0s)

• Duolingo is a $14 billion company that has seen significant growth, with its value doubling in the past six months and hitting all-time highs in usage and market cap (0s).
• The "streaks" feature is the most impactful feature in Duolingo, with over 9 million users maintaining a year-plus streak, and has been the biggest growth driver for the company (5s).
• Duolingo's primary focus is on helping users build habits around language learning, with the biggest problem to solve being getting users to come back the next day (16s).
• The company has run over 600 experiments on the streaks feature in the last four years, with a key takeaway being the importance of testing everything (31s).
• One successful experiment involved changing the call-to-action (CTA) from "continue" to "commit to my goal," which resulted in a significant win (42s).
• The streaks feature has taught the company a lot about human psychology and what motivates and demotivates people, particularly in the context of mobile games and language learning (48s).
• Jackson Shuttleworth, Group Product Manager at Duolingo, leads the retention team and has spent a significant amount of time studying the streaks feature and its impact on user behavior (1m9s).
• The company has set up a robust infrastructure for experimentation, allowing them to run multiple tests and find big wins (2m12s).
• The conversation with Jackson Shuttleworth provides valuable insights into the history of the streaks feature, its biggest wins and wrong turns, and what the team has learned about what works and doesn't work with a streaks mechanic (1m59s).
• This episode is brought to you by Pendo, the only all-in-one product experience platform for any type of application, which helps users uncover what's really happening inside their product and turn those insights into action (3m45s).
• Pendo is built around product analytics, allowing users to see what their users are actually doing in their apps, so that they can optimize their experience (4m11s).
• Pendo also lets users deploy in-app guides that lead users through the actions that matter most, integrates user feedback, and offers session replays to visualize user sessions (4m20s).
• Over 10,000 companies use Pendo today, and users can visit pendo.io/Lenny to create their free Pendo account and start building better experiences across every corner of their product (4m38s).
• Pendo also offers a lineup of free certification courses led by top product experts, designed to help users grow and advance in their career (4m51s).
• This episode is also brought to you by Vanta, a market-leading trust management platform that helps companies continuously monitor compliance, reporting, and tracking risk, and automate evidence collection, unify risk management, and streamline security reviews (5m29s).
• Vanta's platform supports compliance for SOC 2, ISO 27001, HIPPA, and more, and users can save hours by completing security questionnaires with Vanta AI (5m24s).
• Thousands of global companies use Vanta to automate evidence collection, unify risk management, and streamline security reviews, and users can get $1,000 off Vanta when they go to vanta.com/plenny (5m44s).

The impact of streaks on Duolingo’s success (6m0s)

• Duolingo streaks is a feature that tracks the number of consecutive days a user has completed a lesson on the language learning app, with the option to freeze the streak for flexibility (6m7s).
• The feature is simple in theory but has been layered with challenges, goal setting, rewards, and social features, making it a rich feature for the app (6m41s).
• Duolingo streaks have had a significant impact on the app's success and growth, with many companies trying to replicate the feature (7m8s).
• The streak feature is considered the biggest feature at Duolingo, aside from the lessons, and is a great engagement hack that can be applied to any app (7m19s).
• The streak feature is powerful in retaining users, with over 9 million users having a year-long streak, and is a key driver of daily active users (DAUs) (8m23s).
• Duolingo's notifications, which often reference the streak, are also highly effective in engaging users and driving DAUs (7m57s).
• The app's focus on organic growth, including retaining existing users, has contributed to its success, with the streak feature playing a crucial role in this strategy (9m35s).
• Duolingo's valuation has reached $4 billion, with the company's stock hitting all-time highs, and the streak feature is considered a key factor in this growth (8m58s).

The origin and evolution of streaks (9m58s)

• Duolingo's streak feature was part of the app from its launch, with the initial version allowing users to set an XP-based goal for themselves, such as earning 10 or 50 experience points, and extending their streak by achieving that goal (10m5s).
• The original streak system was connected to the user's language learning goal, but it had a drawback: users could lose their streak if they set a goal that was too difficult to achieve, even if they used the app multiple times a day (11m4s).
• An experiment was run to move the streak system from XP-based to a simpler one-lesson-per-day system, which made it easier for users to extend their streak and proved to be a huge driver of engagement (11m27s).
• The one-lesson-per-day system was found to be still meaningful and connected to the user's goal, as the unit of use for Duolingo is doing a lesson, and it shows that users are engaging with the app (12m7s).
• The simplicity of the one-lesson-per-day system is also an advantage, as it is easy for users to understand, and it avoids the complexity of having multiple goals or objectives for the feature (12m31s).
• Experience points (XP) are a key part of Duolingo's system, and users earn them by completing lessons and achieving certain milestones, which drives features such as leaderboards (12m47s).
• Duolingo initially had a leaderboard system where users competed with others to earn XP, but the streak feature is no longer driven by this system (13m4s).
• The company focuses on engagement, as users won't learn if they don't come back to the app, and the easiest way not to learn on Duolingo is not to come back the next day (13m29s).
• As an engagement PM, the goal is to make the app retentive, giving users the opportunity to engage with learning features, and if the experience is dumbed down, users won't care and won't learn (14m5s).
• The streak feature works best when it's part of an app that users care about, and if users don't come back to Duolingo, they won't learn (14m18s).
• The company tracks the impact of changes to the streak feature on the learning experience, ensuring that modifications don't hurt learning and that there's a balance between engagement and learning (14m33s).
• Lise, a Duolingo team member, tweeted that the company optimizes for engagement, as users won't learn if they're not coming back to the app, which aligns with the engagement PM's perspective (13m22s).

Key experiments and insights (14m50s)

• Duolingo's approach to product development involves a strong "test it" philosophy, where they prefer to test different ideas rather than debating them for days, resulting in over 600 experiments on the streak feature in the last four years (17m7s).
• The first version of Duolingo's streak feature transitioned from requiring users to complete a full lesson to earn XP to only needing to complete one lesson, which led to key lessons and insights, as well as wrong turns along the way (14m52s).
• A subsequent experiment tested making it even easier to extend the streak by requiring only one exercise in a lesson, but the results showed that daily active users (DAU) did not increase, and the feature was capturing the least engaged users (15m23s).
• The experiment revealed that the unit of measure for the streak was unclear, and users were not motivated to continue using the app after completing just one question (15m55s).
• The experience taught the importance of considering the type of user being solved for and the level of commitment required to build a habit, as well as the need to be careful when running experiments that may impact a large number of users (16m22s).
• The company's approach to experimentation involves testing a wide range of ideas, from major changes to small tweaks, and being willing to shut down experiments that do not yield positive results, with about half of their experiments being shut down (17m58s).
• The company has a high success rate for experiments, with many companies only having a 20% positive experiment rate, and they tend to shut down neutral experiments to avoid adding cognitive load and complexity to the app (18m9s).
• Neutral experiments are only shipped if there is real conviction around the idea and it provides a new platform to build on top of, and it's recommended to build such features as part of the V1 to ensure the hypothesis around the roadmap has play (18m41s).
• The company focuses on the zero to 7-day user experience, running more experiments to get users to go from a zero to 7-day streak, as data shows that once users reach 7 days, the law of version kicks in and retention increases significantly (19m23s).
• The company found that going from a 1 to 2-day streak is a huge jump in retention, and it decreases slightly until day 7, after which it flattens out, but still increases with longer streaks (19m40s).
• A feature called "streak goal" was introduced, which goals users to hit a certain streak length, and it was found to be powerful in user psychology, starting with a simple version and iteratively adding to it over time (20m21s).
• The company uses a process of testing simple features and iterating on them, rather than designing complex features for V1, which has led to over 600 experiments on streaks (20m55s).
• The idea for the "streak goal" feature was partially inspired by a learning from the monetization team, which found that certain pieces of copy worked well in encouraging users to subscribe to the premium service (21m2s).
• Duolingo initially used a message telling users they were seven times more likely to finish a course if they had a 30-day streak, which was a huge win in indicating an outcome and encouraging users to set goals (21m35s).
• The company then experimented with different streak lengths, testing 14 days and 50 days, and found that they appealed to different users, leading to the realization that they needed to be more thoughtful about who they gave these options to (22m14s).
• An experiment was conducted where users were given the option to opt out of a 30-day streak goal, and if they chose to opt out, they were given an easier goal, which was a good win in capturing users before they said no (22m34s).
• Adding an opt-out button was also tested separately and found to be a huge win, as it allowed users to make an intentional decision about their goals, even though it had no impact on the rest of the feature (23m7s).
• The learning from this experiment was that giving users optionality and intentionality in their goal-setting was a huge insight, leading to the development of a goal-setting feature where users could choose between different goals (23m47s).
• A final learning was that pre-selecting a harder goal for users did not lead to better retention, and instead, allowing users to choose their own goals was more effective (24m18s).
• The goal for users is to maintain a streak, with the act of selecting a specific number of days being a key factor in engagement, particularly around the 14-day mark (24m35s).

User psychology and engagement strategies (24m38s)

• Duolingo's experiments and tests have provided valuable insights into human psychology and motivation, particularly for people who want to learn languages on their phone (24m38s).
• The company's approach to testing involves looking for successful ideas and then experimenting further to see how far they can be taken, often drawing inspiration from other successful apps and games (25m6s).
• Duolingo runs a high number of experiments per capita compared to other companies, and there is a strong culture of cross-sharing insights and ideas between teams, such as monetization and retention (25m35s).
• The company rarely goes into an experiment without a strong hypothesis, often based on previous experiments or successful strategies from other apps and games (25m56s).
• Games like Royal Match and Pokémon Trading Card Game are used as inspiration for engagement strategies, and Duolingo looks at what works in these games and tries to apply it to their own app (26m21s).
• The company's approach to experimentation is not random, but rather based on a hypothesis that has a chance of being true, and the strength of the hypothesis determines the scope of the experiment (26m41s).
• Duolingo has a strong infrastructure for copy testing, and the company believes that copy testing should be done constantly, as small changes in copy can have a significant impact on user engagement (26m56s).
• A successful example of copy testing is the change from "Continue" to "Commit to my goal" on the goal screen, which led to a significant increase in user engagement (27m21s).
• Copy changes are considered cheap and easy to implement, and can be translated into multiple languages, making them a low-risk way to experiment and improve the app (27m44s).
• Luis, the reviewer of all changes and experiments at Duolingo, often encourages testing and experimentation, even if he is skeptical of the outcome, and this approach has led to many successful discoveries (28m3s).

Duolingo’s product review structure (28m36s)

• Duolingo's founder, Lis, reviews all changes to the product to ensure consistency and cohesiveness, which helps prevent the product from becoming a "monstrosity" due to constant experimentation and micro-improvements (28m36s).
• The company has a product review structure in place, which includes the head of product design, product management leaders, and Lis, to maintain a high product bar and prevent the product from becoming inconsistent (29m6s).
• Product managers regularly review features, such as the streak feature, on a quarterly basis to assess progress, identify areas for improvement, and ensure alignment with the company's overall strategy (29m28s).
• Having a clear strategy for a feature is crucial to prevent it from becoming stuck in a local maximum, and AB tests should be done with long-term goals in mind rather than just seeking short-term gains (29m50s).
• Eventually, features may need to be reset and simplified to allow for further improvement and layering on of new features, which can be challenging but is an important process (30m3s).
• Duolingo has an advantage over other companies in that users are motivated to learn a language and are more likely to return to the app, making it easier to implement features like streaks (30m43s).
• The company's focus on education and learning sets it apart from other products, including mobile games, which may use streaks differently due to different user motivations (31m1s).
• Companies can use streaks to give users a sense of accomplishment and motivation, especially when the progress is hard to track on a day-to-day basis, by framing the streak in a way that makes users feel good about it (31m48s).
• The effectiveness of streaks is contingent on companies figuring out how to make them work for their specific product or service (31m50s).
• Streaks work particularly well for Duolingo because language learning is a years-long process, and it's hard to see day-to-day progress, making it difficult for users to track their improvement (32m6s).
• Duolingo uses streaks to show users their progress in a more tangible way, even if they can't quantify the exact amount of progress made, such as "you now know 1% more Spanish" (32m37s).
• The streak feature helps to contextualize progress in a way that makes more sense and feels more tangible to users, especially when the progress is sensed or felt over a longer term (32m55s).
• The concept of streaks can be applied to other companies and products beyond language learning, as long as they can find a way to make it work for their specific use case (33m3s).

Designing for clarity and adaptability (33m7s)

• Duolingo's streak feature started with an XP idea and evolved into a lesson-based system, with the goal of making it simpler and more engaging for users (33m16s).
• The team added streak goals, allowing users to commit to a certain goal streak, and found that users often don't understand how the streak feature works, including how streak freezes work (33m46s).
• To address this, the team ran experiments to make the feature more comprehensible, including adding simple copy changes to explain the streak feature (34m21s).
• One successful experiment involved adding a simple explanation of the streak feature in eight words, which resulted in a significant increase in user retention (34m40s).
• The actual copy used in the experiment was "Start a day to extend your streak, but miss a day and it resets" (35m30s).
• The team considers a "massive win" to be an increase of over 10,000 daily active users (DAUs), and this experiment achieved that (35m52s).
• Duolingo's primary metric for measuring success is current user retention rate (CER), and the team has written about the importance of this metric in a popular newsletter post (36m1s).
• The team's goal is to make the streak feature easily understandable to users, as this is critical for driving user retention and increasing DAUs (34m15s).
• Duolingo's growth is measured by the Northstar metric, which is the current user retention rate, focusing on getting non-new or resurrected users to come back the next day (36m24s).
• The retention team, led by Jackson Shuttleworth, focuses on driving this metric, and the streak feature is the most effective at doing so (36m48s).
• A simple copy change in the streak feature resulted in a significant gain, driving 10,000 incremental daily active users (DAUs) (37m15s).
• Duolingo measures both incremental impact and absolute numbers of new DAUs driven by an experiment, and also looks at retention rates over time, such as day one, seven, and 14 (37m27s).
• The team prefers to look at absolute DAU numbers as it provides an easier comparison, but also considers pitfalls such as recency bias and novelty bias (37m50s).
• The team experiments extensively to avoid creating a product that nobody wants to use in the long term, and has decided not to ship some positive experiments that didn't align with their goals (38m30s).
• One example of such an experiment was adding XP to the UI, which showed a user's progress and earnings, but was ultimately not shipped despite winning, as it didn't align with Duolingo's design philosophy of simplicity (38m55s).
• The lesson screen in the app is a crucial part of the user experience, focusing on learning, and adding other elements can be distracting for users (39m38s).
• Initially, the streak feature was added as an easy engagement win, but it was later realized that it might get in the way and make the lesson experience less simple, leading to the decision to pause and potentially shut it down (40m10s).
• Over time, the beliefs about the lesson experience have changed, and the addition of XP-based mechanics and fun features might make a streak feature work, but it wasn't the case at the time (40m40s).
• The team operates with multiple PMs and teams wanting to show impact, which can lead to the desire to ship features that might not be the best for the user experience (40m59s).
• The streak feature uses a flame iconography, which is a metaphor for a retention mechanic, but this metaphor might not resonate with users from different cultural contexts or stages of life (41m26s).
• User research in India showed that the idea of keeping a flame lit to show commitment did not resonate with users, highlighting the importance of global user research to understand how users experience features (41m51s).
• The screen design for the streak feature was redesigned to focus on the user's progress, using an odometer animation to show the number of days increasing, rather than just a flame icon (42m22s).
• When designing a streak feature, it's essential to prioritize clarity and not get too caught up in creating a beautiful story at the expense of comprehensibility, as form should follow function (42m45s).
• Clarity is crucial and doesn't have to come at the expense of delight; it's possible to make the experience exciting and celebratory while still being clear (43m19s).
• The goal for users who are starting their streak experience is to ensure they understand how the feature works, and the design should communicate the point of the feature and how it works (44m0s).
• Making the streak feature look like a calendar helps users realize it's a daily mechanic, and every element on the screen should be used to communicate the feature's purpose (44m15s).
• Providing flexibility along the journey, such as streak freezes, is a significant opportunity for improvement and a big win (45m45s).
• The Duolingo team uses a holistic approach to screen design, considering every element's role in communicating the feature's purpose and how it works (44m29s).
• The concept of flexibility versus perfection and celebrating regardless is core to how Duolingo thinks about the streak feature, with the goal of not breaking the user's streak if they miss a day, but rather encouraging them to come back and continue using the app (46m16s).
• The idea is to allow users to have flexibility and not be too hard on themselves if they miss a day, but also to encourage them to use the app every day if possible (46m35s).
• The flexibility aspect of the streak feature has been the biggest change from a mechanic perspective, and it is a key part of Duolingo's strategy for the streak (46m57s).
• The streak is designed to be a "ben, not break" feature, meaning that it is intended to be a positive motivator for users, rather than a source of stress or anxiety (46m34s).
• The ultimate goal of the streak feature is to encourage users to use the app consistently, while also being understanding and flexible when they miss a day (46m49s).

Developing the Streak Freeze feature (46m59s)

• One of the earliest experiments conducted was to test the impact of allowing users to have either two or three streak freezes, as opposed to the initial limit of one, and the results showed a significant increase in daily active user retention and return rates (46m59s).
• The experiment revealed that two streak freezes worked better than one, but three streak freezes were no better than two, indicating that there is a limit to the amount of flexibility that is beneficial for maintaining a streak (47m40s).
• The results suggested that while more flexibility can encourage users to return to a streak after a longer period of inactivity, too much flexibility can lead to users taking more time off and eventually losing their streak (48m1s).
• The ideal amount of flexibility for maintaining a streak is still unknown and has been the subject of hundreds of experiments, but it is clear that more flexibility is needed when a user is starting their streak again (48m31s).
• One of the biggest wins from the streak freeze experiments was giving new users two streak freezes when they start a new streak, which made it harder for them to lose their streak initially and increased retention rates (49m0s).
• The data suggests that users need to reach a streak of at least seven days for it to "lock in" and become a habit, and giving users more flexibility at the beginning of their streak can help them reach this milestone (49m43s).
• Conversely, once users have established a long streak, it is generally better not to give them too much flexibility, as they are more invested in maintaining their streak and are more likely to make an effort to keep it going (49m53s).
• The key is to find the right balance between flexibility and motivation, and to condition users to maintain their streak without taking unnecessary days off (50m40s).

Balancing monetization and retention (50m47s)

• Duolingo allows users to buy a streak freeze with gems, which is a feature that can be used to maintain a streak, and the company is currently experimenting with this feature to balance monetization and retention (50m47s).
• The role of monetization and retention in the streak feature is a topic of tension, as making it harder to buy streak freezes can lead to fewer people buying gems, but making it too easy can cheapen the streak (51m20s).
• The company has recently introduced a free option for getting back a lost streak, called "earn back," which requires users to complete a certain number of lessons within a window of time to regain their streak (52m0s).
• The "earn back" feature has been a retention winner, as it makes users feel like they've earned their streak back, and it hasn't cheapened the streak because users have done something to deserve it (52m30s).
• The company is careful not to cross the line and make the streak meaningless by giving out too many streak freezes, and they have a team member, Antonia, who is the "keeper of the sanctity of the streak" and helps to ensure that the streak remains meaningful (53m10s).
• It's possible to get engagement wins by cheapening the streak, but it's essential to hold the line at some point and not make it too easy to extend or maintain a streak, as this can lead to users not caring about their streak anymore (53m20s).
• Having someone invested in the sanctity of the streak, like Antonia, is crucial to ensure that the company doesn't go too far and make the streak meaningless, which would be an "extinction-level event" for the company (53m52s).

Notification strategies (54m8s)

• Push notifications can be a challenge for companies to determine how much is too much, as everyone tends to think that sending another notification is fine, but having a person act as a gatekeeper can help set boundaries and policies for notification limits (54m9s).
• Setting a budget cap for the number of notifications sent can be an effective policy, but creating policies for certain features, like streaks, can be harder and may require relying on judgment and feel (54m32s).
• Duolingo sends two notifications related to the core streak each day: a practice reminder 23.5 hours after the user last practiced, and a streak saver at 10 pm if the user hasn't extended their streak (55m11s).
• The practice reminder time was previously customizable, but Duolingo found that the best indicator of when a user should practice was when they practiced the day before, and 23.5 hours later was the most effective time to send the reminder (55m57s).
• The streak saver notification at 10 pm may seem spammy, but users tend to perceive it as a positive reminder to extend their streak, which is a feature they attach positive emotions to (56m20s).
• A thoughtful notification strategy should consider the feature it's tied to and how users perceive it, allowing for more effective and less spammy notifications (56m58s).
• Duolingo's late-night message has been highly impactful and well-received by users, who often appreciate the reminder to extend their streak before it's too late (57m11s).
• The effectiveness of Duolingo's notification strategy is evident in the stories of users extending their streaks at the last minute, often in unusual or humorous situations (57m51s).

The Perfect Streak feature (58m15s)

• The "Perfect Streak" feature is a simple yet powerful tool that celebrates users' perfection by making their streak look gold and their progress bar on the calendar appear nicer if they don't use a streak freeze for a few days (58m30s).
• This feature serves as a reward for users who maintain their streak without using a streak freeze, and it also helps identify users who aren't relying on streak freezes (59m11s).
• The Perfect Streak feature has no tangible rewards, but it provides a sense of accomplishment and encourages users to maintain their streaks (58m47s).
• Users who don't use streak freezes can maintain a perfect streak, and this feature has become important to many users, with some even reporting bugs when they lose their perfect streak (59m20s).
• The Perfect Streak feature is a counterweight to the flexibility offered by streak freezes, pulling users back into maintaining their streaks (1h0m7s).
• In cases where users lose their streak due to technical issues or other reasons, there are ways to contact the support team, and the company is considering implementing a self-service feature to restore lost streaks (1h0m21s).
• The company prioritizes helping users maintain their streaks, and they would rather provide a way to restore lost streaks than have users give up on their progress (1h0m37s).

Enhancing the user experience (1h0m40s)

• Enhancing the user experience involves making the product feel amazing without relying on copy or features, but rather through animation, haptics, and sound effects, which can make users feel good and celebrate their progress, such as extending their streak in Duolingo (1h0m43s).
• The use of haptics, such as phone vibrations, has been tested and implemented to create a cool haptic pattern when users extend their streak, making the experience more enjoyable and engaging (1h1m16s).
• The goal of these enhancements is to make users pause and enjoy the moment, which can increase their connection to their streak and encourage them to come back the next day (1h2m2s).
• Animation and haptics can be used to celebrate users' progress and make them feel good, which can lead to a stronger connection to their streak and increased engagement (1h2m13s).
• The design of haptic experiences has evolved from being handled by product designers or engineers to being outsourced to a specialized animation studio, Hobs in Detroit, which now handles motion design work and haptics (1h2m49s).
• The role of a haptics designer is unique and requires a specific skill set, making it challenging to hire for this position, as it involves a combination of sound effects, motion design, and technical expertise (1h3m7s).
• Introducing too many features or concepts too early in the user experience can be overwhelming and may lead to a negative impact, so it's essential to balance the introduction of new features with the user's cognitive load (1h4m28s).
• The concept of a "perfect streak" is introduced after users have hit seven days, and the cognitive load of additional streak features is carefully managed to avoid overwhelming users (1h4m11s).

Team operations and experimentation (1h4m47s)

• Duolingo cares a lot about metrics, with most teams being metric-based, and the primary focus is on driving key metrics such as Customer Retention Rate (CRR) and Daily Active Users (DAU) (1h5m14s).
• Having a strong focus on a specific metric helps teams prioritize their work and ensure they are working on the highest ROI tasks, rather than getting lost in trying to make a feature better (1h5m50s).
• Teams are structured around a metric or goal outcome, rather than owning a specific feature, which allows them to work on various products to achieve their goal (1h5m59s).
• The Retention team owns the streak feature, but other teams can work on it if it helps their goals, and there is a culture of soft ownership, where teams are not against others making changes to their features (1h6m6s).
• The Retention team has a roadmap for the streak feature, but other teams can pitch ideas and make changes to it if it aligns with their goals and metrics (1h7m58s).
• The focus on metrics and goal outcomes helps teams prioritize their backlog and gives them clear marching orders, which is beneficial for prioritizing work (1h7m2s).
• Other teams, such as the team focused on increasing time spent on Duolingo, can work on features like leaderboards, and the Retention team can pitch ideas to them if it aligns with their goals (1h6m41s).
• The approach of having teams work towards a specific goal, rather than owning a feature, is similar to how Airbnb worked, where teams were given a goal and could work on any product to achieve it (1h7m9s).
• Multiple quarters of road maps are created for the streak feature, and other teams are allowed to contribute to it, but their plans must align with the existing plans for the streak, ensuring a cohesive approach (1h7m59s).
• Knowledge sharing is essential in this process, where teams share their thoughts, what has worked, and what hasn't, to influence hypothesis and decision-making (1h8m20s).
• A process-oriented approach is crucial when running multiple experiments, and using tools like Jira can help with automation and organization, allowing engineers to work efficiently (1h8m46s).
• A good process involves creating detailed road maps, lining up implementation, and ensuring that design and engineering bandwidth are aligned to roll out the next feature iteration quickly (1h9m22s).
• Planning months in advance is necessary for feature iterations, and losing cycles can result in lost opportunities, so it's essential to manage engineering, design, and product road maps effectively (1h9m57s).
• Mapping dependencies across functions is vital, and while some information is stored in Google Docs, Jira is the primary tool for managing the process (1h10m20s).
• The team resists the urge to create a big V1 and instead focuses on stripping away non-essential features to test the core hypothesis and ship a minimal version first (1h10m45s).
• Antonio, the team lead, emphasizes the importance of being thoughtful about which experiments to run, when to run them, and how to set up the next one to achieve success (1h8m41s).
• When designing features, it's easy to add elements that make them successful, but it's often more effective to simplify the feature and focus on the core hypothesis, making it easier to ship, design, and gain approval (1h11m8s).
• By simplifying features and iterating on them, it's possible to move faster, gain confidence in the hypotheses, and drop parts of the feature that don't work (1h11m54s).
• To operate like Duolingo, it's essential to map all the levers that drive the business, identifying the key metrics that drive growth and daily active users (1h12m27s).
• By focusing on the most connected metrics to growth, such as the streak, it's possible to find opportunities to improve and drive growth long-term (1h12m42s).
• The approach involves exploring and poking at different ideas, coming up with hypotheses and strategies, and trying various approaches to move the key metric (1h13m9s).
• There's always more opportunity to improve, and it's essential to keep working on and iterating on features, as evidenced by the continued improvements to the streak feature (1h13m24s).
• Even after years of testing and improvement, there's still much work to be done, and every quarter brings new opportunities to ship wins and improvements to the streak (1h14m2s).
• The key to success lies in having a strong hypothesis and a clear future strategy, which guides the development and iteration process (1h14m18s).
• Building streaks into a product requires intent and a clear goal in mind, otherwise, it may lead to local maxima, and there is still much to be explored and tried to achieve high confidence in the approach (1h14m24s).
• Streaks are an engagement hack that can make an app more retentive, but it is essential to understand that loss aversion is a human thing that works on humans, and almost every app can be made more retentive (1h14m55s).
• If an app is not something users want to use every day, focusing on streaks will only distract from what really matters, which is making the app enjoyable and providing value to users (1h15m19s).
• It is crucial to have the core loop of the app figured out, giving value to users, and making it something they want to come back to every day before focusing on streaks (1h15m43s).
• Resist the temptation to go too hard down the path of streaks if the core value of the app is not yet providing value to users, as it will likely cause more problems (1h15m55s).
• A streak will not solve problems if people don't care about the core value being provided, and it may even cause more problems if the app isn't engaging (1h16m4s).
• The reason users care about streaks is because the app developers care about streaks, and it is essential to give streaks the visibility they deserve to make them top of mind for users (1h16m37s).
• Duolingo's approach to streaks, such as displaying a big streak screen after every session and mentioning streaks in messages, demonstrates how to effectively make streaks a priority and encourage user engagement (1h16m43s).
• If an app is going to build a streak, it's essential to make it visible and not hide it in a corner, as users are unlikely to care about it if they don't see it (1h17m1s).
• When designing features like Duolingo streaks, it's essential to consider the impact and create a hierarchy of importance to guide users' attention (1h17m33s).
• Effective design cues, such as visual effects and push notifications, can draw users' attention to important features like streaks and encourage engagement (1h17m45s).
• Duolingo's approach to feature development involves extensive testing to determine what works, with a philosophy of "tested first" and a willingness to try new things (1h18m5s).
• To improve or introduce a streak feature, it's recommended to avoid overthinking the philosophy behind it and instead focus on testing hypotheses and understanding user psychology (1h18m16s).
• The most effective way to determine what users respond to is through testing with users, rather than spending time discussing ideas in a conference room (1h18m39s).
• A key recommendation for investing in a streak feature is to focus on figuring out what works through user testing, rather than trying to perfect the feature on the first try (1h18m48s).

Who can benefit from streaks (1h18m57s)

• Most companies can benefit from a streak feature, as it encourages users to return to the app or service regularly, regardless of the specific use case, as long as there is some degree of frequency in usage (1h19m9s).
• The implementation of a streak feature should be tailored to the specific use case of the users, such as coming back every day during a specific season or a certain number of times a month (1h19m17s).
• Companies should identify their ideal usage pattern and build their streak feature around it, as seen in Peloton's weekly streaks, which are more achievable for users than a daily streak (1h19m51s).
• A streak feature can be highly effective in driving growth and revenue, as seen in Duolingo's success, where the streak feature is considered one of the most contributing factors to the company's success (1h20m29s).
• The key to a successful streak feature is to understand the users' needs and usage patterns, and to design the feature accordingly, rather than simply implementing a generic streak feature (1h20m10s).

Lightning round (1h21m0s)

• Jackson Shuttleworth recommends a book titled "A Guide to Midwest Conversation" as a humorous primer on how Midwesterners communicate, highlighting that they often don't say what they mean. (1h21m9s)
• He also recommends "Fate is the Hunter," a memoir of an early commercial airline pilot, which shares fascinating stories about the challenges of flying in the past. (1h21m43s)
• Jackson Shuttleworth enjoys watching the TV show "Bluey" with his kids and recently finished the latest season of "Emily in Paris," which he finds enjoyable despite not being high-brow television. (1h22m38s)
• He is learning French on Duolingo and appreciates being able to understand the French spoken in "Emily in Paris." (1h23m8s)
• Jackson Shuttleworth's mother-in-law is German, and he finds the cultural differences interesting, noting that Germans tend to be direct in their communication, unlike Midwesterners. (1h21m36s)
• He recently discovered a product called Ladder Innovations, which allows one leg of the ladder to extend further than the other, making it safer for him to use on his sloping house. (1h23m41s)
• A favorite life motto is "you miss 100% of the shots that you don't take," which emphasizes the importance of trying things even if the possibility of success is not 100%, as it allows for learning along the way (1h24m29s).
• Duolingo has many traditions, including a weird tradition at every retention standup where the last person to go counts down and everyone tries to clap at the same time, and recently, they've added saying "Yeehaw" in unison afterwards (1h24m55s).
• This tradition started during the pandemic when the team went remote and has continued for four years, becoming an important element of team culture and tradition (1h24m58s).
• The tradition is minor but has become something the team loves and looks forward to, and it's an example of how PMs can find ways to have fun and do something ridiculous (1h25m45s).
• Jackson Shuttleworth can be found online on LinkedIn, and he's open to hearing from people who want to learn more about Duolingo or reach out (1h26m35s).
• Duolingo is hiring product managers, including for Jackson's team, which focuses on thinking about streaks, and interested individuals can apply on the company's website (1h26m45s).
• Jackson is also interested in hearing from companies that have implemented streaks in different ways and learning from their experiences (1h27m7s).
• He used to catalog every streak he found but has found it hard to keep up as the feature has become more popular, and he welcomes interesting streak insights from others (1h27m26s).
• Jackson Shuttleworth, Group PM of the Retention Team, is being congratulated for the impact his team has had, specifically with Duolingo streaks, which is considered a significant achievement for many PMs and teams (1h27m43s).
• The conversation is wrapping up, with appreciation expressed for Jackson's participation and the value of the discussion (1h27m58s).
• The host, Lenny, is thanking the audience for listening and inviting them to subscribe to the show on various platforms, such as Apple Podcast, Spotify, or their favorite podcast app (1h28m6s).
• Listeners are also encouraged to give the show a rating or leave a review to help other listeners find the podcast, and are directed to lenp podcast.com for past episodes and more information about the show (1h28m10s).