The Work

Bling had no central data infrastructure. Reporting was split between Mixpanel and spreadsheets, metrics meant different things to different teams, and a lot of the work was manual. Nobody had a clean view of subscriptions, revenue, or product performance in one place.

I built the company's first proper data foundation using ClickHouse as the central warehouse. Python automations and API integrations replaced the manual reporting work. Event tracking and subscription data were brought into a consistent model, and we agreed on a shared set of KPI definitions across marketing, product, and leadership.

heycar ran across several European markets, but each one reported differently. Metrics were defined differently by country, the data models didn't match, and leadership had no single view of performance. Comparing numbers across markets meant a lot of manual reconciliation.

I standardised the data models and reporting structure across all markets using dbt and Snowflake, agreeing on shared metric definitions that worked at every level. Local teams kept their own operational reporting, and a centralised layer was built on top for cross-market and executive use.

Neonail's marketing data was scattered across platforms with no single view. GA4 tracking had gaps and GTM was inconsistently set up, so key parts of the customer journey — landing pages, product pages, checkout — weren't being tracked properly. Spend decisions were being made without the full picture.

I connected all performance marketing channels into a single data warehouse and rebuilt GTM and GA4 from the ground up to track the full journey: landing pages, product interactions, and the checkout funnel. Attribution was tied to actual revenue, and conversion data was fed back into the ad platforms to improve targeting.

deinhandy’s attribution wasn’t accurately connecting marketing spend to sales outcomes. Reporting was partly manual, and the team didn’t have a reliable view of which campaigns were actually working. Optimisation was harder than it needed to be.

I built a custom attribution model linking marketing spend to real sales outcomes, and integrated offline conversions into Google Ads to improve how campaigns were optimised. Reporting and monitoring were automated with Python, and dashboards were built for both the marketing and business teams. When the company was acquired, I helped manage the transition to new BI infrastructure.