Here’s a question: when was the last time you actually read your organisation’s data strategy? Not skimmed it in a board pack. Read it.

If it was written before 2024, it almost certainly doesn’t account for how your team is actually using data now. And that gap between documented strategy and daily reality is costing you money.

The old playbook

Most data strategies from the early 2020s followed a predictable arc. Step one: consolidate everything into a data lake or warehouse. Step two: build dashboards. Step three: hire a data team. Step four: vaguely gesture at “advanced analytics” and “machine learning” as future aspirations.

It wasn’t bad advice at the time. But the world has moved on.

The businesses that followed this playbook often ended up with beautifully organised data that nobody uses. Dashboards that haven’t been opened since the week they launched. A data team spending 80% of their time on maintenance and 20% on anything resembling insight.

What’s changed

Three things have shifted since those strategies were written.

AI tools need different data than dashboards do. Traditional BI wants clean, structured, aggregated data. AI models often work better with raw, granular, messy data. If your entire data infrastructure is optimised for the first use case, you’re going to struggle with the second.

Data quality matters more than data quantity. The old thinking was “collect everything, figure out what’s useful later.” Now, with AI models trained on your data making actual decisions, garbage in means garbage out with much higher stakes. A wrong dashboard number is annoying. An AI-generated customer recommendation based on bad data is a liability.

Privacy regulations have caught up. The Australian Privacy Act reforms mean you can’t just hoard personal data indefinitely. Your strategy needs to account for data minimisation, consent management, and deletion workflows. These aren’t afterthoughts — they’re core infrastructure.

What a modern data strategy looks like

Forget the 40-page document nobody reads. A useful data strategy in 2026 answers five questions:

1. What decisions are we trying to improve? Start with outcomes, not technology. If you can’t name the specific business decisions your data should improve, you don’t have a strategy — you have a storage bill.

2. Where does our most valuable data actually live? Hint: it’s probably not all in your warehouse. Customer conversations, support tickets, sales call recordings, and even Slack threads contain insights that never make it into structured databases.

3. Who needs access, and how quickly? Real-time dashboards sound impressive, but most business decisions don’t need real-time data. A daily refresh is fine for 90% of use cases. Over-engineering data freshness is a common and expensive mistake.

4. How are we measuring data quality? Not just “is the data there?” but “is it accurate, complete, and timely?” According to Gartner, organisations estimate the average cost of poor data quality at $12.9 million per year. Even small businesses lose thousands to duplicate records, outdated contacts, and inconsistent formats.

5. What’s our plan for AI readiness? This doesn’t mean building an AI platform. It means ensuring your data is documented, accessible via APIs, and governed well enough that you could feed it to an AI tool without causing a compliance disaster.

The people problem

Technology is the easy part. The hard part is getting humans to change how they work with data.

Most data strategies fail because they’re written by the data team and ignored by everyone else. A strategy that requires every sales rep to log calls in a specific format will fail if the sales team wasn’t involved in designing it.

The best data strategies we’ve seen are co-authored. The data team provides the technical framework. The business teams define what they actually need. Finance sets the budget constraints. Legal sets the compliance boundaries.

It’s slower. It’s messier. It works.

Start small

You don’t need to rewrite everything at once. Pick one business process — say, customer onboarding — and map the data flow end to end. Where does data enter the system? Where does it get stuck? Where does it fall out?

Fix that one process. Then move to the next.

A data strategy that improves one thing per quarter is infinitely more valuable than a comprehensive document that improves nothing.

The bottom line

Your data strategy isn’t a technology document. It’s a business document that happens to involve technology. If the people running your business can’t explain it in plain language, it’s not working.

Rewrite it. Keep it short. Make it useful. And for the love of everything, stop putting “data lake” in the executive summary like it means something to your CFO.