How Pratilipi used data and insights to fuel rapid growth

Within its product and engineering teams, Pratilipi set up micro teams, which focus on one specific problem for a month or two at a time. Each team has access to user data in Amplitude, with autonomy to run experiments and put new iterations out with confidence.

“80% of the time we are working on experiments and trying to understand what might work. We develop it, and then we look at the user behavior in Amplitude to understand how people are interacting with it,” says Shally.

One of the teams is responsible for optimizing the onboarding experience—from when a user first signs up, until they’ve read their first piece of content. The team is continuously running experiments across this entire journey, using specific cohorts set up in Amplitude. Analyzing the data, the team discovered that weekly retention is 64% higher for users who finish reading their first piece of content within the first two hours of launching the app. So now they’re encouraging this same behavior in other new users, which has improved their first read completion within a day by 5%.

These experiments are typically run across a small subset of users, and once Pratilipi is confident with the results, the engineering team builds it for scale. When the company launches any new feature in the product, it carefully monitors the core product to ensure it’s not inadvertently breaking things or creating a bad experience for users. Using Amplitude Insight, the company monitors the behavior of roughly two million daily users and gets alerts if there are any anomalies.

Working to hit its growth target of 40 million monthly active users in 2021, Pratilipi’s teams use data to prioritize projects that will have the greatest impact on that metric.

“Unlike enterprise solutions where they are crystal clear on five features in a given release, we have more like 10,000 things we want to try that build up over six months,” Shally said. “If we didn’t keep a check on them, we’d be overrun with tech debt.”

Using historical data from Amplitude, Pratilipi tracks how each new feature is performing, noting changes in usage over time. Likewise, Pathfinder shows them how cohorts have behaved over a specified period of time, and helps them clearly identify what triggers specific users to come back to a certain feature. Even features that have performed very well as MVPs don’t always work well after they’re rolled out at scale and need to be deprioritized.

For example, Pratilipi launched a Reader Dashboard feature with a shareable overview of the content consumed by each of its users. The reaction was positive at launch, but in a three-month study Pratilipi saw use of the dashboard decline. Since this feature wasn’t moving any topline metrics and was burdensome on the backend system, Patalipi killed it, freeing resources for other projects.

“These are important decisions for us to make,” said Shally. “Is the feature a high frequency use in their life? Is it helping our engagement or retention numbers? If not, let’s not waste our time on it.”

Data-driven culture runs throughout the entire organization at Pratilipi. The marketing team, for instance, meticulously tracks retention numbers in Amplitude. They’ve found that the type of new users coming in from different campaigns and advertising channels makes a difference in whether those users will come back to the product.

By integrating their attribution tool—Branch.io—with Amplitude, the marketing team can see each user cohort coming from its campaigns and track retention. This allows the team to compare campaigns across all channels, not only in terms of cost-per-install, but also how many of the new users will remain on the product over a period of time. They can then double down on the best performing campaigns.

As Pratilipi starts to monetize its platform this year, its teams are confident in how the numbers are going to look six months down the line. They can see it in the data.