Contrary to what you might have heard, we are stuck with Data Warehouses, whether we like them or not.

Let’s not get stuck in the differences between a “Data Lake”, “Warehouse”, “Silo”, “Data Intelligence Platform” or “Kevin” (Real name of a system out there). If it (1) merges; (2) scrubbed data; (3) in a form that makes analysis easy; and (4) with high performance; it’s a data warehouse.

And data warehouses aren’t going away anytime soon, because the problems they solve haven’t gone away. We still need a place to store data from a number of different systems that we can represent and reuse relatively easily for reports, business intelligence, dashboards and the like. And, even when these tools are offered “In the cloud”, as some sort of SaaS solution, they are still Data Warehouses, smelling just as sweet as anything called a “warehouse” is going to smell.

Why do we need them?

It pretty much always comes down to the same incident—An executive in the Business is struggling for insights or information to make a decision, and the existing enterprise systems can’t quite provide needed information at a glance, or, worse, they can but their reports disagree.

The discussion after this incident generally sounds something like this:

• The IT team demonstrates that they’re collecting the right data, but it isn’t in the right order, timeframe or format, and doesn’t live inside a single system.
• The Business executive asks searching questions about why, after all the company’s investments in CRM, ERP, reporting tools and more, they still can’t easily answer simple-sounding questions.
• Conversations are remembered about how these systems were supposed to provide fantastic reporting and even better dashboards.
• Frustration ensues. But eventually, blamestorming fatigue sets in and everyone involved figures out that the company needs a coherent analytics strategy, supported by IT in the form of … some form of data warehouse.
• Ideally, this will all be accompanied by a more collaborative working relationship between business executive management and the IT organization.

Occasionally, there is resistance. Most often it’s rooted in a blame-oriented business culture – the need to spend time, money, and opportunity costs must be someone’s fault. And as a general rule, whenever there’s fault to be assigned, IT is the logical and convenient scapegoat.

Cut this off at the knees early and often. Explain that there’s no fault to be assigned. New requirements need new solutions, and new solutions aren’t free.  Let’s talk about some of the questions that you may get, and how to deal with them head on.

• This seems like a big project.

Bite the bullet early and acknowledge that data warehousing projects, whether built around highly structured “snowflake” data models or data-dump-based HDFS data lakes, are never small and simple. But they don’t have to be unmanageable. There are ways to stage implementations so delivery happens at a satisfactory cadence. The complete analytics roadmap can get big, but the tools, technologies, and practices needed to support the effort are better than ever.

• Aren’t we doing extra reporting work?

Not really, you are just doing the right work to create effective outcomes. The CRM/HRMS/whatever systems by themselves aren’t designed to give you the right intelligence at the right time – not through any inherent deficiencies of being a transactional management system, but because no one system has all the data needed to support the desired analyses.

• I don’t really want Procurement/HR/Supply chain (Read- Other executives) to see my data.

Your response: “That’s a great point. As part of the roadmap we’ll definitely want to make sure the right mechanisms and processes are in place to make sure only the right people in the right roles have access to the right data.”

Next week: Specific tips and tricks for getting organized for success.