Fund the outcome, not the platform

A board does not buy a medallion architecture. It buys control, faster decisions and numbers it can trust. Here is how to put a data platform investment the right way up.

Most data platform business cases are built upside down. They start with the architecture, move to the platform, then the pipelines, then the reports, and somewhere near the end they mention the benefit. By then the board has stopped reading. The number they are being asked to approve has no outcome attached to it, so it looks like cost, not investment. We build the case the other way up, and it changes how a board hears it.

The case most boards see, and why it fails

A finance director asks whether to spend a large sum on a new data platform. The technical answer is yes, because the business needs a governed estate, Fabric capacity, pipelines and semantic models. That answer is true and it is the wrong one. It hands the board a shopping list and asks them to trust that value falls out of the far end. The better answer names the outcome the money buys.

Name the outcome the money buys

Each of these is a number a board can weigh. None of them is a pipeline.

• Control. Can revenue and gross margin be trusted, reconciled and traced back to source, or does finance rebuild them by hand every month before the board sees them?
• Decision speed. Can management have reliable numbers on day two after month end instead of day seven? Five days a month is sixty days a year of running the business on last month’s picture.
• Operating efficiency. Can the hundreds of hours spent on manual reconciliation be removed, or redirected to work that actually needs a human?
• Working capital. Can better visibility of receivables, stock or supplier positions release cash the business already has but cannot see?
• Financial performance. Can better data surface revenue leakage, margin erosion or unprofitable customers that are invisible in a spreadsheet?

Do you need to spend the whole sum?

An outcome-led case forces a second question into the open, and it is the one that builds trust with a board. Perhaps a quarter of the sum solves the control problem on its own. Perhaps half automates the highest-value reporting and the rest can wait until that has paid for itself. A case that only knows how to justify the full figure has not been thought through. A case that can defend each slice against an outcome is one a board can approve with confidence, in stages, as the value proves out.

Why we lead with a fixed-price discovery

The most valuable work, the modelling and the AI, is only as good as the estate beneath it, and that estate is rarely ready on day one. Discovery establishes what is true and what must become true before the rest is committed. It turns a large, vague number into a sequence of smaller ones, each tied to a result.

What we would do

If you are being asked to approve a data platform this quarter, turn the case up the right way. Name the business problem. Attach the financial or operational impact. Then, and only then, put the investment against it and state the outcome you will measure. If your data estate is heading toward a decision like this and you would rather size it against outcomes than architecture, talk to us at hello@hoptonanalytics.com.