Net revenue retention for PostHog's Data Warehouse fell from 118% to 104% over two quarters. What is going on?

Introduction

PostHog's Data Warehouse (now part of its broader "data stack" with Sources, Destinations and batch exports) lets teams sync data from Stripe, Postgres, HubSpot, Zendesk and other sources, then query it with SQL alongside product events. Customers pay per synced row beyond a free tier of 1 million rows a month. Net revenue retention (NRR) measures how much revenue a cohort of existing Data Warehouse customers generates now compared with a year earlier, including expansion, contraction and churn. A fall from 118% to 104% over two quarters means existing customers are expanding much less, or contracting and churning more. Because Data Warehouse is what lets PostHog answer revenue questions ("which features do our highest-paying customers use?"), weakness here threatens the all-in-one strategy. This is a hard question because NRR is a lagging, composite metric, and the cause could be pricing, product, customer mix or competition.

Step 1

Clarifying Questions (3 minutes)

• How is NRR calculated for Data Warehouse?

• Why it matters: Usage-based products have volatile monthly revenue.
• Hypothetical answer: Trailing 12-month revenue from customers who paid for Data Warehouse a year ago, divided by their revenue a year ago, measured monthly.
• Impact: Changes in billing rules show up directly.

• Did anything change in pricing or billing?

• Why it matters: Pricing changes are a common cause of revenue shifts.
• Hypothetical answer: Two quarters ago, incremental syncs became the default for large tables, and historical syncs became an add-on.
• Impact: Fewer synced rows could reduce revenue even with stable value.

• Are customers churning from PostHog entirely or only from Data Warehouse?

• Why it matters: Separates a product problem from a platform problem.
• Hypothetical answer: Overall PostHog NRR is stable.
• Impact: The problem is specific to Data Warehouse.

• Has the positioning or target user changed?

• Why it matters: The handbook notes research into data engineers' needs after Data Warehouse was repositioned.
• Hypothetical answer: Yes, it was repositioned from "a warehouse for product engineers" towards a broader data stack.
• Impact: Customer mix and expectations may have shifted.