Data warehouse as a service: What it solves, and what it doesn't

Table of Contents

• What is a data warehouse as a service?
• What do you get with a data warehouse as a service?
• What else do you need?
• What is a full-stack data platform and how is it different?

What is a data warehouse as a service?

A data warehouse as a service is a cloud data warehouse that enables you to load, model and query data with SQL. Snowflake, BigQuery and Redshift are the usual examples.

But are you looking for a data warehouse, or a full-stack data platform?

What do you get with a data warehouse as a service?

The data warehouse gives you:

1. Centralized storage - one place to put data from all of your systems
2. Access & permissions - a way to provision access, apply permissions (who can access what) and query it quickly
3. Ability to overcome standard reporting - As a company expands, it inevitably encounters the limits of standard reporting functionality (often referred to as "native reporting") of business apps e.g. Salesforce, NetSuite, HubSpot, Marketo, etc. Having your data in a warehouse gives you the flexibility to combine, pivot, filter and otherwise manipulate the data to suit your analytical and operational data needs.
4. Ability to analyze data without ramifications - production databases like MySQL and Postgres should not be queried directly as this could impact the performance of the applications that are reliant on them. For example, if you run a complex query against the Postgres database that powers your company's software, that could impact the end user experience!

So for all the reasons stated above, pulling the data out of your systems/databases and into a purpose-built environment for running analysis and building your data models is the sensible thing to do.

But there's a catch: you still need to get the data out of your sources and into the warehouse.

What else do you need?

Quite a bit, actually.

The data warehouse is empty and basically useless until you build out the supporting infrastructure.

1. An ETL tool to sync data from your CRM, marketing automation, ad platforms, finance system, product database, etc.
2. Data models (i.e. SQL) that join those sources into things like pipeline by campaign, cost per opportunity or multi-touch attribution
3. A business intelligence or visualization tool
4. A reverse ETL tool, if you want a score or segment to land back in your CRM or marketing automation platform
5. One or more data engineers to set up and maintain all of the above

That last one is the big one. It's the reason having a data warehouse doesn't always mean that data is accessible.

A data warehouse is a good option if:

• you have massive data sets - billions or trillions of rows - which should be compressed, aggregated and analyzed
• you have the budget and the personnel (i.e. a team of data engineers) to build and maintain this infrastructure

But this is not the reality for most companies. So, what about a full-stack data platform?

What is a full-stack data platform and how is it different?

A full-stack data platform gives you everything you get in a data warehouse + a lot more.

You get centralized storage, access & permissions, complete flexibility to overcome standard reporting, and the ability to analyze data without ramifications.

But it also gives you automated, scheduled extraction to keep your data fresh and up-to-date, data modeling to store versioned queries, so everyone shares the same system of intelligence, the ability to create visualizations from your data to help get your point across to business stakeholders, and much more.

If you want the whole thing, not just the warehouse, you want a full-stack data platform. And the good news is that you can try Datachef for free for 30 days to see if it's right for you. Connect your data (sandbox is fine) and get a feel for how Datachef can help you make better decisions, faster, without having to build and deploy 6 or more different data tools.