This blog post is Human-Centered Content: Written by humans for humans.
It’s a question we will bravely ask every client if we find out they are still utilizing self-hosted hardware. It’s usually met with a pause. Not because the answer is complicated, but because most organizations have never had a reason to say their approach out loud to someone outside of their company.
We’ll commonly get a mix of, “Our data still lives where it always has,” “It’s spread across on-premises servers,” “It’s a maze of Excel files and whatever tool happened to be handy when someone needed an answer from the data team,” etc.
I saw this similar story myself earlier this year. For years, a client utilized a mix of Tableau Prep flows and Excel files that pretty much held together the reporting backbone everyone relied on. And it worked — in the sense that reports got built and questions got answered. But every new question relied on heavy data discovery, and every change to the existing ETL added one more thing someone had to remember to update by hand.
Then one day, a member of our team popped the question that we enjoy asking our clients when appropriate: “Remind me, why aren’t you utilizing a CDW?”
It’s a fair question, and once it’s asked, it’s impossible to ignore. Excel files and a series of complex Tableau prep flows are not exactly a true data strategy. They are the accumulated result of a reporting team’s analysis needs (and usually a lack of time or resources to rethink all of the modern data stack advantages for the organization).
Even worse, business objects could get created using more legacy tools to structure the outputs of the data patchwork, and a company becomes susceptible to getting locked in time.
What Changed Once the Question Got Asked
Asking “why no CDW” opened the door to a full data transformation discussion, not just the potential need for a new tool. As the team mapped out where data actually lived, and what was duplicated, along with what was stale and what nobody trusted anymore, lights started going off. Honestly, the hardest part of the entire effort of migrating data to a CDW isn’t discussing technological needs but just understanding the current state of the data with complete clarity.
Part of that clarity meant being honest about what needs to go. The old technology, along with the business objects, reports and prep flows built up over years that would no longer be needed. Take those as the first to get sunset.
In its place is a sourced data layer approach. Data lands in the CDW in its rawest form first, before any business logic gets applied to it. From there, it then gets organized into clean, structured tables inside the warehouse itself, rather than scattered across desktop tools and spreadsheets. That shift means the transformation logic lives in one place, is repeatable and is visible to anyone who needs to use it.
What came out the other side was a single source of truth. Not a slogan, but a real, specific outcome: One place where numbers are refreshed on a schedule, where the definition of “revenue” or “active customer” doesn’t change depending on which spreadsheet you opened, and where the answer to a question is the same whether the CEO asks it or an analyst does.
That kind of consistency is the backbone of good, modern data decision-making. It’s also durable and agile over time. A cloud data warehouse built well today should still be the trusted foundation a decade from now, not something the next team has to tear out and redo. Oh, and hey, AI is knocking. How easily can you let them in to a walled-off maze? A worthwhile CDW already has the power of AI in its capabilities.
The Usual Suspects
For many organizations, the shortlist of cloud data warehouses tends to look at the same players: Snowflake, Databricks and Microsoft Fabric show up consistently, alongside Google’s BigQuery and AWS’s Redshift. Each has real strengths and real trade-offs, and the right choice depends on the existing tech stack, the team’s skill set and where the organization wants to be in five years, not just where it needs to be next quarter.
None of these platforms are magic. Moving to the cloud doesn’t fix bad data on its own, and it doesn’t replace the work of deciding what “single source of truth” actually means for a specific business. What it does is remove the ceiling. On-premises systems and manual file-based processes have a hard limit on how much they can scale, how fast they can refresh and how many people can trust them at once. Cloud platforms don’t have that ceiling.
(For reference, see: Enabling Snowflake Cortex AI in Governed, High-Control Environments)
All top-tier CDWs provide enterprise-grade security environments that support GDPR, HIPAA, CCPA and PCI DSS. The old bottlenecks have been removed with many fully compliant providers to choose from.
InterWorks will always share its best deployment configurations and strategy on CDW projects. Some of the biggest challenges are ensuring the data stays secure, within network and are seamlessly a part of the user validation infrastructure. We’ve worked with many Fortune 100’s and large public sector organizations who need to operate under strict corporate standards and government requirements.
The Question Is Worth Asking Yourself (Soon)
If your organization is still running on the foundation of Excel, legacy exports, macros and manual refreshes that have gotten you this far, the question is worth sitting down with us for a consultative chat. We’re open. We’re honest. We’re tool-agnostic. A true data partner will look into your ecosystem, make sense of it and help with a path to the best CDW setup.
On-premises tools are not inherently bad, but the gap between “it works” and “it’s built to last” are the reason we’re so passionate about this topic. If you’re ready to find out what your own answer might be to the, “Why no CDW?” question, we’re glad to help you work through it.
