Start with one decision the business already makes

A new governance program is often introduced as a framework: councils, principles, a glossary tool, and a charter. The framework is not wrong. It is too large for the first month, and it is not attached to a decision anyone is waiting on.

Pick a decision the business already makes. The monthly forecast. A filing that has to be explained to a regulator. The list of customers who receive a discount. Ask which numbers in that decision people recheck by hand, export to a side file, or refuse to use. That distrust is the scope.

Put the first month in order

Work through the decision in this sequence.

• Name the decision and the person who makes it.
• List the data sets that decision actually uses.
• Name an owner and a steward for each set that would change the decision if it were wrong.
• Write down the terms those people already argue about, and give each term one meaning.
• Write one quality rule for the field that causes the most rework.

Stop there. A second domain can wait until this loop has run once: an issue was raised, someone owned it, and it closed. A roadmap that promises every domain in the company before that has happened is a poster.

What you should be able to show

At the end of that first pass, the evidence is ordinary and specific. A short list of data sets. Names next to them. A handful of definitions. One rule that can be tested. A place where an open break stays visible. That package is a governance program. The framework can be written to fit it, instead of the other way around.