How Compare the Market cut report delivery from 6 weeks to under 1 by retiring most of its dashboards

At dbt Summit 2026, Emanuela Ciotti, head of enterprise analytics at price comparison company Compare the Market, shared a framework built from 7 years of working with dbt across 3 different organizations.

She opened with a warning about the AI push most analytics teams are feeling: AI does not validate the data it consumes, it trusts it, so a shaky data foundation doesn't get exposed by an AI layer on top of it, it gets amplified by it. That set up the core problem she called the "reporting factory": a BI team operating as a service desk, SQL logic duplicated across reports and tools, and metrics defined differently from one dashboard to the next.

An internal audit of Compare the Market's own Power BI estate found 433 total reports, 41% of them (179) already abandoned, each one a separate, one-off answer to a separate question with no shared foundation underneath.

Ciotti's framework breaks the resulting transformation into 3 stages. Discovering covers the shift from analyst to analytics engineer, where version control, modular design, and testing get introduced before any new tooling.

Scaling addresses what happens when a single monorepo with more than 4,000 interdependent models becomes a structural liability, particularly when part of the staging layer depends on Google Sheets seeds that occasionally broke financial pipelines right before month-end reporting. Her team's fix was a multi-repo architecture aligned to domain ownership, drawing on dbt Mesh principles to give each business tribe ownership of its own models.

Her team is now in the transforming stage : redefining what the analytics function actually produces, from dashboards to tested and documented data models, and positioning that data layer as the foundation AI needs to be trustworthy.

The results she shared were concrete: time to deliver a new report dropped from 6 weeks down to under 1 week, test coverage passed 75%, and model reuse climbed to 3.0x.

This is fundamentally an operating-model change, not just a stack change. A team only moves from service desk to platform when what it hands the business changes from a dashboard to a governed model, and that requires designing ownership in from the start rather than assuming it will emerge on its own.

Watch every keynote and breakout from dbt Summit 2026 on demand. Watch dbt Summit 2026 on demand →