Taxfix is on a mission to simplify tax filing and help people save time and money. The company’s tax app acts as a smart assistant asking users simple, personalized questions without any forms to maximize their tax refunds. As users start answering questions, Taxfix surfaces only relevant follow-up questions and hides questions that don’t apply. The app also accepts photos of payslips (so there’s no need to file forms) and submits filings directly to the tax office.
Founded in Berlin in 2016, Taxfix launched their app in 2017. In those few short years, they’ve helped collect over €420 million for their users–many of whom had never had the confidence or patience to file a tax return before.
Avoiding data bottlenecks and inconsistencies
Taxfix needed a single source of truth for their customer data that could be integrated with their existing systems quickly. As their team scaled, they knew they had to watch out for a potential “data swamp,” which would create bottlenecks, slow down projects, and cause data inconsistencies across teams and tools.
To make informed decisions, the team wanted to ensure a high level of confidence in their underlying data. So before jumping to a solution, the engineering team created a comprehensive list of features for their future data platform:
Distributed ownership of data: product teams should be able to add, delete or change an event and be able to see the change in their reports.
Quality pyramid: quality issues should be flagged quickly and as close as possible to the data producers.
Optimized for writes: Optimizing for writes will combat data silos and integrate data into the data lake faster.
Codified knowledge: The platform should automate naming conventions, schema-as-a-code, and monitoring/alerting.
Self-service possible through automation: Moving quickly and avoiding bottlenecks is a top priority.
Easy to discover (data catalog): The team needs clean data with clear meaning to build actionable insights.
Easy to extract & distribute (APIs, SQL, webhooks, streaming, etc): Data use should be as democratized as possible across our teams.
Implementing Twilio Segment in one month
When weighing whether to build or buy their dream data platform, Taxfix’s decision came down to ROI and time to value. The team decided that engineers could better spend their time building and improving the product rather than building infrastructure. Their engineers are experts in serving their customers, so acting on user feedback was ultimately far more valuable than integrating and maintaining their internal platform.
Instead of spending years building an in-house ingestion pipeline, Taxfix implemented Twilio Segment in one month.
They integrated Twilio Segment Protocols with their data catalog to define events through a schema, which ensured validation and quickly identified any tracking errors. Meanwhile, it was easy to subsequently send that data downstream to the tools used across the business. The team’s CRM tool, Intercom, integrated with the product in a few days, as did their A/B testing solution.
Twilio Segment’s raw data queue can also be stored in Snowflake, making it easy for the Taxfix team to reprocess data if there is a mistake of re-send data to a new source.
Optimizing the product and making an impact with customers
With Twilio Segment, the Taxfix team deepened their confidence in their data and decision making. Team members now have the context they need to use data more effectively and Twilio Segment is there to guide and protect against mistakes along the way.
With a data lake built to scale and a renewed trust in their data pipeline, the Taxfix team has more time to optimize their product and make a real impact on their customers.
Interested in hearing more about how Segment can help you?
Data instrumentation has been significantly simplified, dramatically reducing engineering costs on building and maintaining data pipelines.