Managed data warehouse: What does "managed" actually cover?

Table of Contents

• What does "managed" mean?
• What is needed to answer business questions?
• Why do teams with a warehouse still use spreadsheets?
• So ... what do you need?

What does "managed" mean?

When Snowflake, BigQuery or Redshift call themselves a managed data warehouse, they mean the warehouse runs itself. You don't tune it or patch it. You load data, write SQL and pay for what you use.

That's genuinely useful. But "managed" stops at the edge of the warehouse. Getting data in, making sense of it and putting it in front of people is still on you. (I went through the full list in Data warehouse as a service: What it solves, and what it doesn't.)

The easiest way to see the gap is to follow one question all the way through.

What is needed to answer business questions?

Say your CMO asks: "Which campaigns sourced pipeline last quarter, and what did each one cost?"

The first challenge is understanding where the data resides.

It may exist in one place - your CRM - or it may span multiple systems (e.g. Salesforce and Hubspot/Marketo, etc.)

Then you need to determine how to interpret "sourced" -- is there a definition you can use based on fields/designations within the CRM (e.g. a direct link between Campaign and Opportunity) or is an inferred attribution model needed for such a question?

A different business question: "How many qualified leads were generated last week? Were they engaged" Here again, we need to define what "qualified" means -- this could mean leads with certain job titles, locations, etc. We also need to define "engaged" - this could mean that a phone call was placed after the sales rep received the lead.

No matter what the question is, shared definitions are what make this sort of measurement possible. These don't come pre-packaged with your data warehouse, of course. You need to define these, store them somewhere, and make sure your team (and possibly other teams) have access to the same definition. This ensures that everyone has the same view into key business metrics and there is no deviation in reporting.

Why do teams with a warehouse still use spreadsheets?

Because setting up everything I described above is complicated.

I've seen plenty of companies with access to a data warehouse where marketing still exports CSVs every month.

The warehouse isn't the problem. Nobody has the time or comfort level to build and maintain stuff in Snowflake. As is so often the case, you don't know what you don't know - until you start building.

So ... what do you need?

A managed warehouse is the right call if you have data engineers who can support each department. They'll need the granular control over each component of the data infrastructure, and the warehouse gives a solid base.

If the people asking questions are in marketing, sales, finance, product - and nobody has time to build and babysit a stack, you want more of it managed: the syncing, the modeling and the reporting, not just the storage. In other words, a full-stack data platform rather than a warehouse. That's what we built Datachef to do, and our team runs it for you.

It doesn't have to be either/or. Plenty of teams keep their data warehouse and data infrastructure for large-scale data storage and processing (i.e. petabytes of proprietary data that are compressed and aggregated) work and carve out the business reporting to a platform like Datachef that is more accessible in a self-service fashion to business teams.

Want more than the warehouse managed? Watch a demo, or try Datachef free.