Every BI team eventually faces the same question from their users: “Where does this data actually come from?”

When users can’t answer that, they get frustrated, and their work is delayed when they feel they must stop and check someone else’s work. And (a BI Dev’s worst nightmare), they’ll think in the back of their minds “This would be easier if I did it in excel”.

One51 Consulting faced this problem at a large property investment and management organisation. We decided to solve it holistically, in a way that would uplift and enhance the report development process.

Client context: scaling reporting in a fast-growing data ecosystem

Within the last 2 years, our client has launched an extensive data platform and reporting suite that is still growing every month. With this scale and velocity come the usual complications:

• Multiple source systems, each with its own identifiers for the same entities. One entity might have three different “friendly names” across systems, or two entities in one system might roll into a single entity in another.
• Different departments anchored on different primary systems. Each team wanted reassurance that the numbers in front of them came from their source of truth.
• Inconsistent terminology and business definitions between teams, even when they were ostensibly talking about the same thing.
• Reports expected yesterday by which we mean that reports were expected to be live almost as soon as data landed in the data platform.

What we needed to achieve

1. Align the look and feel across the entire reporting suite.
2. Capture the lineage of every field and metric in every report.
3. Streamline the end-to-end development of new reports.
4. Expose that lineage to users in a way that felt intuitive and was sustainable for the BI team to maintain in new and existing reports.

Documentation initiatives die when they become extra work, so we needed the practice to live inside the BA and developer workflow, not bolted onto the side of it.

 

Designing documentation that lives inside the workflow (not beside it)

1. Document lineage at the visual level, in a standard template.
For each report, Business Analysts completed an Excel workbook capturing every visual, every metric and dimensional context in the visual. Then include its business definition and the source-system lineage all the way down to the specific front-end field or fields driving the calculation. This could be created off the back of the existing wireframing process that BAs undertake with Stakeholders, allowing BI developers to understand the required fields and work with data engineers to ensure that all the required fields are captured from source.

2. Create a BI Glossary semantic model that combines all the standardised glossary files.
We published the glossary model with read access across the organisation, then connected the glossary back into every report, using a semantic model live connection. New and existing reports now carry a direct connection to the glossary, with lineage tagged down to the report > page > visual level.

4. Surface the glossary in-context with a popup on every page.
Users can search by visual name and instantly see the relevant terms, definitions, and source systems without leaving the report they’re looking at.

Trust, speed, and consistency across every report

• A cohesive, brand-aligned reporting suite that looks and feels like one product.
• Business definitions, calculations, and lineage at users’ fingertips, exactly where and when they need them. Trust and transparency went up because the answer to “where does this come from?” is now one click away.
• Real efficiency gains for the BI team. Lineage and definitions are captured as part of the BA process, which means developers no longer hunt through source systems for the right columns. The required fields can be ingested to the reporting platform before report development even begins.
• Less time spent reviewing lineage and answering user queries.
• Model and report design logic is now documented, so domain knowledge and developer expertise are far easier to share across the team.
• Increased uptake across certified reports and models.

From static documentation to living metadata

The documentation is now a living asset.

With Power BI MCP and Copilot in Fabric on the horizon, we’re well placed to take advantage of it. Definitions and synonyms can be inserted into model and report metadata to train Copilot so that the glossary can be the foundation for a genuinely intelligent, conversational reporting experience.

Power BI MCP also makes handling change requests even easier for developers to document. Metadata and changes to visuals can be extracted from project files and fed back into the glossaries.

With lineage, definitions, and context embedded directly into every report, users no longer ask “where does this data come from?”. They explore it confidently in real time, and the BI team finally scales trust as fast as it scales data. If your business needs help with this, please contact our One51 Consulting BI specialists.