How not to create a Data Strategy: A cautionary tale of Tech-first thinking.

One of the less fun parts of this job is reviewing failed data strategies mostly representing thinly disguised investment cases for big technology purchases. You know the ones — glossy decks filled with supplier logos, architecture diagrams that look like train maps, and a budget spreadsheet that reads like a supermarket sweep at a todays cloud provider of choice flagship store. These strategies don’t fail because the technology is bad. They fail because they start with technology. And that, in our opinion, is the cardinal sin of a data strategy. 

Let’s find out why. 

The call of the shiny 

It usually begins with good intentions. A senior leader hears that “Your data is an AI powerhouse,” attends a conference, and comes back convinced that the company needs a data lake, a lakehouse, a mesh, or some other aquatic metaphor. And that’s before we get started on LLMs. Anyway, a task force is assembled, consultants are hired and before anyone has asked what problem we’re trying to solve, there’s a procurement process underway for a platform that promises to do everything short of making great coffee. 

This is not a strategy. It’s a shopping list. 

Technology is seductive. It’s tangible. It comes with demos and dashboards and promises of transformation. But buying technology without a clear understanding of your data needs is like buying a Ferrari before you’ve learned to drive — expensive, impressive, and ultimately useless and costly. 

The mirage of strategic alignment 

Now, to be fair, not every failed strategy starts with a technology catalogue. Some begin with organisational objectives — which sounds promising. But here’s the catch: those objectives are often so high-level, so abstract, that they can’t be operationalised. “Become data-driven.”, “Unleash AI”, “Unlock customer insights” and “Accelerate innovation.” These are aspirations, not strategies. 

Without a clear path to land those goals — through culture, governance, and capability — they remain stuck in the clouds. You can’t build a bridge to a destination you haven’t mapped- if you don’t know where you’re going, you’ll end up somewhere else. So if your organisation doesn’t have the cultural maturity or data operating model to support any or all of these aspirations, even the most well-intentioned strategy will fail to land. 

The missing questions 

A real data strategy starts with questions. What decisions are we trying to improve? What data do we need to make those decisions? Where does that data live? Who owns it? Who trusts it? What’s broken in the current process? What are the most important things we need to fix? Do we have the skills we need? 

These are not glamorous questions. They don’t come with vendor endorsements or Gartner quadrants. But they are the foundation of a strategy that works. 

When you skip these questions, you end up with a solution in search of a problem. You build pipelines that no one uses, dashboards that no one trusts, and governance frameworks that no one follows. You spend millions to create a data ecosystem that’s technically brilliant and strategically irrelevant. 

The cult of the platform 

One of the most common symptoms of tech-first thinking is the cult of the platform. The belief that if we just buy the right tool — the one with the most features, the best integration, the slickest UI — everything else will fall into place. 

Guess what? It doesn’t.  

Platforms are enablers, not saviours. They can accelerate a good strategy, but they can’t compensate for the absence of one. If you don’t know what data matters, who needs it, and how it should be used, no platform will fix that. You’ll just end up with a very expensive sandbox. 

The governance mirage 

Another hallmark of failed strategies is the governance mirage — the illusion that buying a tool with built-in governance features is the same as having a scalable data operating model. 

A real governance capability starts with people and culture. Its foundations are trust, accountability, and a shared understanding of what to do and how to do it. It requires conversations, compromises, and clarity that’s never going to be automated by a tool. You can’t click your way to data accountability. 

When governance is treated as a checkbox in a technology implementation, it becomes performative at best. Policies are written, approved but never enforced. Roles are defined, allocated but never understood or acted upon. Metadata is captured but never trusted. So, when things go wrong — as they inevitably do — no one knows who’s responsible. Too often we start again in the wrong place. 

The people problem 

Perhaps the most tragic consequence of tech-first strategies is the neglect of people. Data is not just a technical asset; it’s a social one. It lives in systems, yes, but it also lives in spreadsheets, emails, conversations, and tacit knowledge. 

If your strategy doesn’t engage the people who produce, consume, and interpret data, it will fail. You need to understand their pain points, their incentives, their habits. You need to build trust, not just pipelines and dashboards. 

Technology can support this, but it can’t replace it. A dashboard doesn’t create insight. A data catalogue doesn’t create understanding. These things come from people — curious, sceptical, informed people — asking good questions then making better decisions. 

A better place to start 

So, having spent a few hundred words trashing technology first approaches, let’s provide some guidance on how to start with your data strategy 

Start with a simple purpose for the strategy to fulfil. What are the strategic goals of the organisation? What decisions drive those goals? What data do those decisions require?  

Then move to people. Who makes those decisions? What do they need? What do they trust? What’s broken? What skills do they need to make best use of a high quality trusted data sources? 

Only then should you think about process. How is data currently collected, stored, shared, and used? Where are the bottlenecks? Where are the risks? What are the priorities A strategy cannot fix everything, but it can fix something so choose wisely! 

And finally — finally — think about technology. What tools can support the processes, empower our people, and serve the strategy’s purpose?

This sequence is not optional. It’s the difference between strategy and shopping. 

Last words 

If you’re reading this and thinking, “But we already bought the platform,” that’s not a disaster. You can still build a strategy, but you’ll need to reverse-engineer it. Start by asking the hard questions. Revisit the purpose and engage the people – that’s all the people not just those in the technology domains. Then be brutally honest about what the technology can and – as importantly - can’t do. 

A good data strategy is not a technology implementation. It’s a business capability. It’s messy, human, imperfect and iterative. It’s not about buying the right tools — it’s about asking the right questions. 

So next time someone hands you a data strategy that looks suspiciously like a supplier brochure, smile politely, and ask: “What problem are we trying to solve here?” 

If they can’t answer — or if the answer is “become data-driven” — you’re not looking at a strategy. You’re looking at a very expensive mistake. 

On that note, check out our Data Tangle research project. Our presentation at Data Decoded was titled “How to fix data issues with no budget and no staff” - it has been compared to “Scrapheap Challenge for Data”!