Executive Summary
Your engineering team has built a good data pipeline, but the CFO and the board won't use the numbers because they don't trust them.
This can undermine the data team's work. The new data stack starts to look like an expensive overhead, while the business keeps relying on spreadsheets.
The issue isn't usually the pipeline itself. It's the gap where raw data needs to be turned into business metrics in a way finance can audit. It's a problem of design, not just engineering, and it's about connecting the two teams.
What happens in the Monday morning meeting
It’s a scene I’ve seen a few times in growing companies. As the CTO, you present a dashboard showing healthy pipeline runs, good uptime, and data delivered on time. Your team has done its job well. The system works.
Then the CFO holds up a printed spreadsheet, looks over their glasses and says, “I appreciate the technical work, but these numbers don’t reconcile with our general ledger. We can’t use this for the board pack.”
The room goes a bit quiet. Your technical success hasn't translated into a business one. The conversation hits a wall, and suddenly there are questions about the whole data project. It's as if you've built an efficient pipeline for crude oil, but the CFO needs petrol, and the refinery is missing.
Why the pipeline numbers don't match the finance spreadsheets
This isn't your engineers' fault; it's a gap in the design. In the push to modernise, it's common for growing companies to invest heavily in warehouses like Snowflake and tools like dbt. They build very capable systems for moving data about. What sometimes gets missed is the final step: turning that raw data into business metrics that are trusted and can be checked.
You've probably moved to the cloud and hired good engineers. But if the underlying process is a bit messy, all you've done is automate the mess. You're just getting the wrong answers faster.
The problem isn't that your dbt models are failing. It's that the official definition of 'Net Revenue' or 'Active User' still lives in a complex Excel formula on the Finance Director's laptop. This is the gap where data trust tends to fall apart.
I see this quite often with companies around Series B or C. They have a great technical setup, but the business rules aren't written down in code. The result is usually the same: the Finance team still exports to Excel because it's the one place they can completely control the sums.
How to connect the engineering pipeline to business logic
The answer isn't to make the pipelines faster or hire another data engineer. It's to start treating the business definitions themselves as a product. This means shifting your CTO data strategy from just the flow of data to its meaning.
This is where you would design and build a governed Semantic Layer. This isn't about buying another tool. It's the patient work of sitting down with the CFO, the head of sales, and the COO to understand and unpick their spreadsheets and manual calculations. You take the rules that live in their heads, or in a spreadsheet, and put them into a central, version-controlled system.
What you end up with is a Single Source of Truth that engineering automates and finance has signed off on. The logic is out in the open, written down, and, crucially, everyone agrees on it.
The process involves people, not just code
You should expect some resistance. Building a proper single source of truth can be difficult because you're asking people to give up the tools and processes they've relied on for years. You're effectively retiring the 'master spreadsheet' that the finance team might have spent the last decade perfecting.
This kind of work is often more about diplomacy than code. It means running workshops, writing things down clearly, and showing people that the new, automated system is more reliable and easier to check than the old manual one. You will have to move slower for a few weeks to move faster for the next few years.
What the board meeting looks like afterwards
When this is done well, the Monday morning meeting changes completely. The CFO presents the board pack directly from the company's BI tool. When a board member questions a metric, the CFO doesn't have to 'check the spreadsheet later'. They can show, right there in the meeting, the exact logic in the code that produced the number.
The arguments tend to stop. Trust is rebuilt. Your data stack stops looking like an expensive technical cost and starts being the reliable engine for the business. Your engineers' hard work is no longer lost in translation. It becomes the foundation for the important decisions the company makes.