Meditopia Transforms Data Analytics With Fivetran and dbt

Key Takeaways:

Dramatically increased conversion rates and grew subscriptions to more than 3 million
Saved the cost of hiring one full-time data engineer for ETL
99.9% up-time and zero maintenance on ELT
Optimized advertising spend decisions
Improved user experience with deeper product insight
Native integration with dbt makes data ready for analysis faster

Data Stack

Pipeline: Fivetran

Source: Snapchat Ads, Apple Search Ads, Google Ads, Facebook Ads,

Destination: Google BigQuery

Business Intelligence: Tableau

We're saving months of work; you can set up 10 different connectors in a week. There's simply no way a human being could do that. We can make decisions much faster because we have been able to cut down the feedback loop with timely and perfect data.

Gustav Lindqvist, Head of Data, Meditopia

Meditopia, a successful startup founded in 2015, is a mental wellness platform that provides over 350 meditations and mindfulness exercises in seven languages to more than three million people. Using a freemium business model, customers are acquired with a trial on the website or via iOS and Android apps that the company looks to transition into paid subscribers. Data analytics and insights are crucial to conversion and producing personalised content that people are willing to pay for.

The Challenge

Drowning in Pools of Data

The business is driven by digital advertising and using data from various platforms – Apple Search, Snapchat, Google, Facebook, etc. – to optimise the effectiveness of the marketing spend. The company had started by building reports from third-party sources without a centralised place to collect, collate and analyse the data. It required a lot of manual effort, downloading CSV files that were copied and pasted into Excel.

Meditopia’s main strategic goal, like most startups, is growth. To meet this objective, it was vital to acquire more insights into the business. The data was scattered across different silos and painstakingly integrated into a data lake -- a pool of raw data that could subsequently be pulled into a data warehouse, Google BigQuery, for processing and analysis.

Gustav Lindqvist, Meditopia’s Head of Data, explained that the company considered two options: building its own pipelines or outsourcing to a specialist. The first option would mean hiring a data engineer, which was a problem because good ones are hard to find and expensive. Lindqvist said,

For us it was a no brainer, we knew we were going to pay a service provider such as Fivetran to do it.

The Solution

Tools for Transformation

One of the reasons Meditopia chose Fivetran was because of its native transformations with dbt, an open source analytics engineering tool that enables transformation models to be run directly from source applications. With Fivetran dbt Transformations, Meditopia could orchestrate the cleansing, transformation and modelling of data sets and make it ready for analysis.

“The less providers you have providing your Virtual Data Pipeline [VDP] tools the better,” said Lindqvist.

It’s very valuable to have Fivetran as one provider who can take care of both steps -- getting data into the data lake through its connectors, and then using dbt to transform the data in the data warehouse.

Meditopia was soon up and running with 50 different models, executing batches twice a day, and syncing them in BigQuery where data is analysed for two different use cases. Marketing is the main one, where Meditopia analyses and optimises spend decisions, making sure the company gets more ‘bang for its buck’ in digital advertising.

The second reason for utilizing Fivetran is product improvement. The company uses Amplitude as their product intelligence platform to help inform the meditation content and get a better understanding of what resonates with subscribers. For a business that runs on a freemium ‘try-before-you-buy’ model, identifying ways to optimise the user experience and convert them to a purchase was business critical.

The Results

Fast Access to Insights With Added Resilience

Dramatically reducing the time it takes to get to insights has been a big benefit of Fivetran. “We're saving months of work,” said Gustav Lindqvist. “You can set up 10 different connectors in a week. There's simply no way a human being could do that. We can make decisions much faster because we have been able to cut down the feedback loop significantly with timely and perfect data.”

Meditopia has improved its conversion rates and grown its subscriber base to over three million users. All of this has been significant, but the real game changer for the data team is how monitoring works right out of the box. According to Lindqvist, you never see any service interruption because Fivetran takes care of all the management and maintenance 24/7/365..

“That’s a life changer,” said Lindqvist. “When a pipeline you built breaks, you have to fix it yourself . I have learned the hard way how frequently it happens and how it always happens at the worst possible time. You have to drop everything and fix it. With Fivetran, their core business is to make sure things work.”

Fivetran removes the risk of concentrating all the knowledge of a customised system on one employee. If that person walks out the door, the business is exposed. Fivetran, in contrast, is easy to use and despite running quietly in the background, makes a massive difference to the business. “You know something is business critical when you realise what would happen without it; that you would have to replace it with something more expensive and less efficient,” said Lindqvist.

About Fivetran: Our fully managed pipelines and automated connectors for databases, applications, events, files and functions help extract value more quickly from data, accelerating the path to becoming a data-driven business and achieving competitive advantage through insight-based decision making.