A Data Strategy on One Page

If you signed off a data strategy this year, try one test.

Name one decision where the number beats the gut, agreed before the meeting. If nobody can name one, I’d say the strategy hasn’t started yet, whatever the deck says.

The usual answers are all true

Is the data clean enough yet? Shouldn’t the platform land first? Isn’t a director paid for judgment? All true. But none of it explains why a meeting still ends on “my feeling is” where the data is perfectly fine. From what I’ve seen, nobody agreed beforehand which number settles the question.

That is a strategy gap, and it is a small one. It does not need a larger programme to close. It needs a page.

Three columns per decision

In July I wrote that value is a named person and a decision they were going to make anyway. Passing today’s test takes the next step: one page that holds three entries for every decision you keep making. The decision. The one number allowed to settle it. The person who owns that number.

Take next week’s staff roster. The demand forecast sets it, and your planning lead owns the forecast. When the planner’s gut says Friday needs two more people and the forecast disagrees, the page already says which one wins, because you agreed on that weeks ago. (The roster, the forecast and the planning lead are an illustration, not a scene from a client.)

That agreement is the strategy. The owner of the number signs up for something real: when the forecast is wrong, the planning lead hears about it first, and fixing it is their job. Which means the business owner of the decision has to accept the number in advance, and the data team has to keep the number honest. Both sides get a name.

A page like this can start with three rows. One filled in, two left empty on purpose, because the empty rows are the invitation to the next two decisions in your organisation.

Why the forecast gets to win

Letting a number overrule a planner sounds cold until you look at the record. In 2000, Grove and colleagues pooled 136 studies, from fields such as medicine and criminology, that pitted a formula against expert judgment. The formula was about 10% more accurate on average, however experienced the experts were. The spread is worth knowing: the formula was clearly better in 33% to 47% of the studies, and judgment was clearly better in 6% to 16%.

That holds for repeatable predictions, like a weekly roster forecast. It does not hold for everything. Opening a branch in a new city is a one-off bet, so I’d leave it off the page. Your gut has plenty to do there. The page covers the decisions you make again and again, and those are exactly the ones where agreeing in advance pays off.

A short list you can check

Gartner surveyed 504 data and analytics leaders in autumn 2024. Of those, 91% call creating an actionable strategy a primary responsibility, and 22% track business impact for most of their use cases. I won’t pretend a page of decisions closes that gap everywhere. It is measurable by design, though: a few weeks in, you can check whether the roster got better. A strategy deck rarely offers that test.

There is a practical consequence for the platform question. Your Fabric or Azure plan gets sized from that page, once it exists. Three recurring decisions, each with one authoritative number and one owner, tell you which data has to be reliable first, how fresh it must be, and who needs access. Without the page, those choices get made by analogy to the last project or by whoever argues loudest.

Where to start this week

Pick one decision your organisation makes on a rhythm: a weekly roster, a monthly forecast, a quarterly pricing review. Ask three questions in a room with the person who makes it and the person who owns the data behind it.

1. Which number should settle this when opinions differ?
2. Who owns that number, and what do they commit to?
3. What would make us change the rule?

Write the answers on one page. Then check the next time the decision comes around. If the room followed the number, you have a strategy with one row in it. If the room did not, you have learned which part of the agreement was missing, which is more than the deck told you.

Which decision in your organisation passes that test today, and who owns its number?

Happy to think this through with you.