A production marketing attribution platform in 11 months

Twistag built Source.app, a marketing attribution platform that unifies data from six marketing platforms into a single source of truth, from concept to production MVP in eleven months. Source.app's founders had watched customers reconcile Google Ads, Meta, HubSpot, Google Analytics, Search Console and Mailchimp by hand and reach contradictory conclusions about the same campaigns. Twistag opened with a six-week technical feasibility phase — validating the data model, integration strategy and event schema before committing engineering — then built separate executive and campaign-manager views on a normalised event architecture, with an AI Analyst layer that flags anomalies proactively.

Marketing leaders face a fragmentation problem that compounds as their tech stack grows. A campaign manager might check Google Ads for search performance, Meta for social ROI, HubSpot for pipeline impact, Google Analytics for site traffic and Search Console for visibility, and still have no coherent narrative of what is working — because each platform counts differently and the numbers contradict each other. The cost is hours lost to manual reconciliation and decisions made on incomplete data. Source.app's founders saw the pattern across their customer base and decided to build the unified layer. The hard part was never the APIs. It was designing a data model and an interface fast enough to validate the idea before market momentum moved.

Twistag structured the engagement in three linked phases, with the consequential decision taken at the start. The first six weeks were technical feasibility — not design, not engineering, but focused validation of the data architecture, the API integration strategy and the event tracking model that would have to support the platform for years. That phase established the shape of the problem before heavy engineering was committed, and it forced the early argument about which data belongs on an executive dashboard versus a campaign manager's view.

Twistag then built the platform across two audiences deliberately: executives get high-level ROI across channels, campaign managers drill into daily performance, bid optimisation and underperforming creative. Separate views cost more design and development time than a single interface, and are the reason the product reads as cohesive rather than overwhelming.

The integration layer connects Google Ads, Meta, HubSpot, Google Analytics, Search Console and Mailchimp, each with its own authentication, rate limits and data model, normalised through a custom event architecture into a common schema. On top sits an AI Analyst that surfaces anomalies proactively rather than waiting to be asked — when a channel's ROI shifts significantly, the system flags it with context. Integrations run on a schedule, pulling fresh data every four hours. Each integration was built against its own authentication model and rate limits rather than a shared abstraction, because the six platforms differ enough that a common connector would have leaked complexity into the schema.

Stack: React, Next.js, AWS, with a custom Node.js service orchestrating the API integrations and running the AI analysis layer. Services: AI-ready data platforms, product engineering, product design, AI integration.

Source.app shipped its production MVP in November, just under the eleven-month target. The platform is live with early enterprise adopters using it as their primary marketing performance dashboard, and campaign teams report spending less time pulling data across platforms and more on strategy and optimisation. The engagement also validated the delivery method: starting with a feasibility phase meant the engineering phase ran without architectural pivots mid-build. Stated plainly, these are structural and qualitative results — Source.app has not published a client-side revenue, cost or efficiency figure, and this page does not claim one.

Committing to a data model before validating API complexity is how multi-month rebuilds start.