Detailed subscription analytics for this AI assistant.
Objective
Build a trusted analytics foundation to drive growth across revenue, marketing, and product.
Obstacle
Misaligned metrics and scattered tools led to slow, unreliable reporting.
Outcome
With a unified data pipeline Fyxer now makes faster, insight-driven decisions.
Background
Fyxer AI is a UK-based platform that provides executive assistant services powered by tech and human expertise. As their user base scaled, so did the need for clear performance visibility. Previously reliant on Stripe, spreadsheets, and Chart Mogul, their reporting lacked consistency and depth. To modernise their analytics, Fyxer partnered with 173tech for a project focused on building a scalable data stack, defining core metrics, and delivering cross-functional dashboards. The initiative spanned finance, marketing, engineering, and customer success, underscoring its strategic priority across the business.
Challenges
Misaligned Metrics: Before the project began, Fyxer had no consistent definitions for key business metrics such as Monthly Recurring Revenue (MRR), Churn, or Average Revenue Per Organisation (ARPO). Different stakeholders referenced different sources, leading to confusion and decision paralysis. This made it difficult to run cohort analyses, measure the impact of discounts and promotions, or assess the effectiveness of customer retention efforts.
Opaque Attribution: Marketing data was siloed, with ad spend recorded in Meta platforms and results only loosely matched to user signups. Attribution logic was unclear, resulting in underutilised performance data and an inability to confidently optimise marketing spend. Without a clear link between campaigns and business outcomes, marketing decisions were made on intuition rather than evidence.
Solution
Data Pipeline: 173tech established a modern analytics stack leveraging Fivetran for automated data ingestion, BigQuery as the data warehouse, and dbt for transformation and modelling. A central GitHub repo was set up with CI/CD pipelines using GitHub Actions to ensure reliable, version-controlled deployments. Metabase was chosen for dashboarding, and access roles were configured to support both technical users and business teams.
Model Development: Key metrics were modelled across multiple domains including MRR, Subscriptions, Marketing, Users, and Organisations. For revenue, 173tech built both gross and net MRR models to reflect discounts and transaction fees. Cohort retention, churn by age and ARPO by cohort were also implemented. For marketing, attribution models were developed by ingesting UTM-tagged signup data, enabling per-channel and per-campaign analysis. A comprehensive dashboard wireframe was signed off early in the project, ensuring that final visuals aligned with business needs.
The Proof Is In The Numbers...
10x
95%
6
Implementation
Core Metrics: Fyxer now operates with a single source of truth for all revenue, subscription, and churn metrics. Finance and leadership teams no longer debate numbers pulled from different systems. Dashboards provide daily updates on key metrics like net MRR, churn rate split by customer tenure, and NRR. This clarity has improved financial forecasting and surfaced previously hidden churn trends.
Accountable Marketing: The inclusion of cold email and paid social attribution data in a unified model means that Fyxer’s marketing team can finally quantify ROI by campaign and platform. Dashboards show cost-per-signup, conversion rates by cohort, and attribution splits by source. This has enabled more confident budget allocations and test-and-learn strategies, particularly useful in a lean and performance-driven startup environment.
Cross-Team Visibility: Custom dashboards were developed for Customer Success and Support, allowing those teams to proactively manage churn risk and respond to service issues more effectively. By automating the generation and refresh of analytics data, the engineering team is no longer burdened with manual exports or ad hoc SQL queries. With well-documented definitions and user-friendly tools, the business now makes faster, more informed decisions.
