How to Make Sure Data Access Doesn’t Kill Your Project

You’re leading a new data initiative. You have a bold vision. Resources? Check. Stakeholder support? Check. You know what to build, what data you need, and where it lives. All you need now is access.

And that one little detail can bring the whole thing crashing down.

80% of data projects never make it to production. Across the dozens of projects I’ve observed, data access has often been the silent killer. It's surprisingly hard—and slow—to get access to data that's already sitting in your organization, managed by teams who know how to access it. But without it, your well-resourced, well-scoped, and well-supported project stalls.

Even with all the initial momentum, if you can't show value quickly, enthusiasm evaporates. By the time you finally get access to that key data source, the project is already deprioritized.

So let's talk strategy. Here are seven practical ways to deal with the most common data access blockers before they sink your initiative.

1. Plan Conservatively for Data Access

Don’t build your milestones around optimistic timelines for data access. You will be wrong. And when everything else is ready but you're still waiting on one Kafka topic or database view, you have nothing to show.

Stakeholders don't want to hear that it's just a quick data hookup away. They want results. So build your plan assuming delays. Add generous buffers. Make sure early milestones don’t depend on data you don’t yet have.

In software engineering, conventional wisdom says that you should multiply timeline estimates by a factor of 2-3. For data access, I would say multiply by a factor of 5-10 depending on the sensitivity of the data. It sounds egregious, but all it takes is one compliance review process to blow up your estimates.

This raises the obvious question: if I need data to deliver value, but I don’t have access to the real data yet—what do I do?

2. Settle for Lower-Grade Data Access

You may want a real-time CDC feed from the core system, but maybe there's already a nightly dump in a data lake you can use. Or an S3 backup you can extract. Or a CSV the team uses for reporting.

It’s often much easier to get access to lower-grade versions of the data—even if they’re outdated, lossy, or incomplete. That’s okay. Use what you can get. Prove value now. You can improve the fidelity, latency, and volume later.

3. Use Samples

Sometimes the data is too sensitive or locked down to get full access early. But you might be able to get sample data. Anonymized test sets. Sanitized exports. Demo data.

Yes, they’re limited. Yes, they won’t cover every edge case. Yes, you’ll have to do some hand-waving. But they’re enough to build workflows, run demos, and give stakeholders a sense of what the real thing would look like.

4. Fake It Till You Make It

If you can’t get samples either—because of politics, workload, or because the data hasn’t even been captured yet—make fake data. Use data generator libraries to mock realistic events, users, and interactions.

No, the data won’t be perfect. It’ll be worse than samples and you’ll have to wave your hands even stronger during the demo. But it gives you something to work with. Something to show. Something that helps people visualize what success could look like. That’s a lot better than Figma mockup.

Oh, and one final word of hard-learned advice: Resist the temptation to write your own data generator. Yes, the dozens of libraries you can readily find on the internet all fall short in one way or another for your use case. But building a better one turns out to be surprisingly hard (“we need to build our own sampling algorithm to get the product to order relationship multiplicity right”) and consumes time and resources you don’t have. Take what’s out there and hack around the sharp edges.

5. Do More with Less

Wouldn't it be great if you could get access to all the sources you need right away? Absolutely. But sometimes it’s possible to deliver value without all the data.

You might delay a specific feature. Or de-emphasize a particular data stream in your initial rollout. The key is to show value early—and if a subset of your data can get you there, go for it. In fact, try to plan for that up front. It will significantly de-risk and de-stress your project.

This buys you time. It also builds credibility, which makes it easier to unlock additional access down the road.

6. Blame the Game, Not the Players

Sometimes the roadblocks are people. Legal has concerns. Security wants audits. The data team doesn’t want to be on the hook. It's frustrating.

But resist the urge to get adversarial. Most people aren't trying to stop you—they're trying to protect their responsibilities. So focus on the incentives. Look for allies. Understand their pressures. De-escalate. Empathize.

Realize that the obstacles are partly necessary to protect the organization and partly a symptom of organizational dysfunction. And everybody is trying their best to play within that ruleset.

Understand for what the game is and what everybody’s position is within the game. Realize that technical arguments are often cover for underlying concerns about roles and responsibilities. So, instead of debating the technicalities of why you need direct DB access, speak to their role and what success would mean for them too.

7. Take the Stone Soup Approach

There’s an old story about a traveler who tells a hungry village he can make soup with just a stone. Curious, the villagers gather. He puts a stone in boiling water and says, “Tastes good! But it would be even better with carrots.” Someone brings carrots. Then potatoes. Then herbs. Eventually, everyone contributes, and they make a rich soup together.

This is how data initiatives often succeed. You don’t asking for all the data access you need. That leads to resistance and debates. But if you show something—even something scrappy—people will help you make it better.

Give them a reason to believe. Demonstrate a compelling partial result. Show how one team's dataset made it more useful. Then point to how another dataset could make it even better. Invite participation. Celebrate contributions. Succeed.

In short, don’t let perfect data access be the enemy of early momentum. Start with what you have, fake what you don’t, and show just enough value to pull more access into reach.

Build the soup. Let them bring the carrots.

And always remember: agility and momentum is your friend. Inertia is your enemy and data access is a top source of inertia. Use these strategies to make sure data access inertia doesn’t kill your project.