13 weeks to a single source of truth
Objective
Build and deliver an automated executive dashboard for Digital Ventures, consolidating revenue, ads, moderation, users, and website data.
Obstacle
Digital Ventures had no governed modelling layer, no single source of truth, and key data disconnected from the warehouse.
Outcome
A fully automated, multi-tab executive dashboard in Metabase, built on validated dbt models in BigQuery.
Background
Digital Ventures operates a number of classified advertising platforms. As the business pushed toward accelerated growth targets, their biggest internal blocker was not strategy, it was data. Leadership could not get a consistent answer to basic questions. Revenue figures differed depending on who you asked. Ad performance was hard to measure. The BI team was overwhelmed with one-off requests.
173tech had already mapped the problem during a prior strategy engagement. The prescription was clear: stop patching the existing setup and build something clean from scratch.
Challenges
No Single Source Of Truth: Digital Ventures had data in ,multiple systems, none of which shared a common identifier or agreed definition of core metrics. Revenue figures differed between Finance, the BI team, and the payment gateways. Ad counts varied depending on which extract you were looking at. There was no modelling layer to standardise KPIs, absorb upstream changes, or give teams a shared foundation to work from.
Data Quality Problems: Many of the decisions that the DV team had made in setup made perfect sense at the time, but as technology had evolved and the business had become more complex, quality issues had accumulated. This included missing data on adverts, revenue and order data. No one had visibility of the scale of these problems until someone tried to model the data end to end.
Solution
A Governed Data Layer: 173tech built a clean, isolated pipeline and dbt modelling layer directly inside Digital Ventures’ own GitHub repository, structured on a medallion architecture. A data dictionary was agreed before any modelling began, locking down metric definitions across revenue, ads, moderation, users, and website performance. Upstream schema changes could now be absorbed without breaking downstream reports, and the BI team had a codebase they could own, extend, and maintain independently.
Custom GA4 Pipeline: GA4’s standard BigQuery connector did not support the custom metrics Digital Ventures needed; specifically contact events tied to ad IDs, user IDs, session source, medium, and channel. 173tech built a custom extraction method to pull this data into the warehouse and model it alongside MySQL data. For the first time, they could see website behaviour: sessions, attribution, and user contacts in the same place as revenue and ad performance.
Impact
Revenue Reconciled Across All Channels: The final revenue model spans multiple payment channels, currencies, sales types, and geographies and was validated against Digital Ventures’ existing reference reports before go-live. The unmatched rate came in below 0.01%, giving Finance and leadership a number they could stake decisions on for the first time.
60% Of Contacts Attributed: Using the custom GA4 pipeline, 173tech was able to tie together contacts back to a verified user ID. For a platform where understanding who contacts whom is central to the business model, that’s a meaningful baseline. The remaining 40% is now clearly flagged and understood, rather than simply missing.
Quality Issues, Fixed During the modelling work, 173tech surfaced missing data which had been distorting numbers for years. Now they are fixed, documented, and handled automatically in the pipeline. Going forward, every model, every pipeline, every definition lives in Digital Ventures’ own GitHub repository and runs on their own infrastructure. 173tech built something DV owns completely, not a dependency they will be paying to maintain.
Creating Value For Digital Ventures..
84 KPIs across Users, Ads, Moderation, Revenue & Product,
Sub 0.01% reconciliation rate on revenue,
And a clean, stable analytics platform for the future.
Get In Touch
Our friendly team are always on hand to answer questions, troubleshoot problems and point you in the right direction.
