Roadmap

Goals

Q4 2026 objectives

Objective: Build the features data engineers expect (Daniel)

• Key Results:
  • Ship SCD2 tables so users can see how a record changed over time.
  • Add governance features that give data teams control over who can see what.
  • Support PrivateLink and the other major cloud providers, so data moves over a private connection.
  • Launch rETL, so users can push PostHog data back into the tools they run on (Tom).

Objective: Improve the surfaces people use to work with their data (Tom)

• Key Results:
  • Find what the MCP is still missing, and improve the parts that already exist.
  • Fix the parts of the UI people still use heavily, and point users to the MCP for the tasks it does better.
  • Make the difference between ClickHouse and the warehouse clear in the UI, so users know what we charge them for.
  • Ship self-driving signals, so the product tells users about problems in their data.

Objective: Make the pipeline cheaper, more accurate and broader (Estefania + Tom)

• Key Results:
  • Decrease the cost of running the pipeline, and improve our margins.
  • Give users a more accurate and more real-time view of their data.
  • Ship the ETL product.
  • Move the billing team off Fivetran and onto our own pipeline.
  • Fill the gaps in the sources we already have, including missing fields and tables.
  • Support Iceberg, Delta Lake and DuckLake for tools outside PostHog.
  • Build better support and debugging tools, so we can diagnose sync problems faster.
  • Add more sources, always and forever.

Objective: Tell people what we built (Lizzie + Tom)

• Key Results:
  • Launch the ETL product, with marketing behind it.
  • Publish engineering blog posts about our work.
  • Explain when to use the CDP, rETL, batch exports and data pipelines, so users pick the right one.
  • Close the gap between data engineers and product engineers in how we talk about the product.
  • Run a marketing series: "data engineering for product engineers".

Objective: Learn what users actually want from us (Anna)

• Key Results:
  • Improve retention. We currently do nothing about customers who churn.
  • Run more customer interviews, because we get too little feedback today.
  • Talk to data engineers too, not only product engineers.
  • Run MCP and UI interviews, to find where each surface falls short.

Handbook

Who are we building for?

Personas

• Primary Personas:
  • Data/Analytics Engineer
    • These are the engineers building and maintaining the whole data stack, and are also responsible for providing the best tooling for their company.
    • They want flexibility and control of every part of the stack to ensure correct reporting downstream.
  • Data Analysts/Product Managers (technical type)
    • They are responsible for providing insights for their teams/company. If a question can be answered by data, they will write complex SQL models to get to the answer.
    • Data analysts and product managers are the power-users of data modeling and business intelligence (BI). They have the desire and the time to go significantly deeper into the data.
• Limited focus:
  • Product engineer
• Not a focus but should be usable by:
  • All other engineers, e.g. they should have access to the data they need, and easily query it