Grippi connected acquisition to product retention
Grippi is a SaaS marketing data platform built for agencies. It pulls client ad and analytics data into one place so agencies can report and act on it faster. The company sells on self-serve trials, so the path from first visit to paid account matters a great deal.
Grippi had no reliable picture of which features drove retention. Marketing data lived in one set of tools and product usage lived in another, and nothing joined the two.
The team could see which campaigns brought in signups. It could not see what those signups did next, where trial accounts stalled, or which channels produced customers who stayed.
Engineering also wanted a second pair of eyes on product and API decisions that affected how data moved through the platform.
Marketing APIs into the warehouse
We built custom marketing APIs that push campaign and attribution data into the warehouse on a schedule. Each source lands in a consistent shape with source, medium and campaign kept intact, so joins downstream stay clean.
Web and product tracking
We implemented GA4 and GTM on the marketing site and PostHog inside the product, sharing identifiers across both. A visitor who signs up keeps the same identity from the first ad click into the app.
Trial and cohort instrumentation
We wrote an event taxonomy for key product actions, trial start, activation steps and conversion to paid. We then built trial-to-paid funnels and retention cohorts in PostHog so product and growth teams could see where accounts stalled.
BigQuery as the single source
All sources were consolidated into BigQuery and modelled into tables for acquisition, usage and revenue. Usage questions now start from one tested model, and the MCP servers read from it too.
MCP servers for conversational queries
We built MCP servers that let AI assistants query the data in plain language. The team and its customers could ask questions of their marketing data without writing SQL or waiting on a report.
Product and API consultation
We worked directly with engineering on product and API design, covering data contracts, naming and how new features should emit events. This kept tracking reliable as the product changed.
- Acquisition channels are now tied to what users do after signup, across every platform.
- Usage questions are answered from a single source of truth in BigQuery.
- The team can see exactly where trial accounts stall in the funnel.
- Feature adoption and retention cohorts are live in PostHog.
- The MCP layer drove a large increase in client acquisition.
Related work
Ad spend matched to real applications, including the offline ones.
We rebuilt tracking on a server-side GA4 setup, brought every source into one warehouse and built attribution that ties spend to completed finance applications.
A global GA4 migration with no break in measurement for any market.
We designed a global analytics schema for Royal Canin, migrated a very large multi-market site to GA4 and piped the raw data to its Azure and Databricks teams.