Our engineering team is deciding which data warehouse to integrate with our system, and the BI tool to interface with it.

Preliminary question - Is it best practice to try and consolidate all data to be analyzed in one location (warehouse) then have the BI tool just interface with that one source to draw insights? Know some BI tools can connect to multiple but not sure if that's a crutch until teams are able to create a single destination for all of their data

Business Requirements

We're looking to create dashboards for each company KPI - with the primary KPI as the highlight of the dashboard, then other downstream metrics that impact it alongside of it
We're looking to sync data across the platforms we work with: Stripe, Twilio, Sendgrid, Salesforce, Facebook Ads, Google Ads, Paypal, Business Amazon account (not AWS)
For the BI tool, we want to be able to share dashboards, connect different API's and databases, have flexible date ranges, and a nice to have is easy to interface with if team members don't know SQL

Current stack

Segment to route user events to Google Adwords, Facebook Ads, Mixpanel, and S3
Mixpanel to analyze web and mobile metrics
Fullstory for enhanced mobile and web visibility
Salesforce as a CRM - majority of our data lies within here

Current thoughts

AWS Redshift seems to be well adopted, integrate with most tools, and we're already building on AWS so it seems to make sense. BigQuery seemed more expensive and Snowflake didn't seem terrible but wasn't in AWS ecosystem
Looker has looked the most impressive on the BI tool side, but open to discussion
We're looking to do this alongside other projects with an in-house engineer and a contractor - we're a bit limited on the technical resources and we're looking to at least get a first pass in and eventually enhance the integration as we have bandwidth

Guidance / advice is appreciated, even if it's only for data warehousing or BI tools specifically (and not both)

Michael Copley

Tools Used

Most Used

By Layer

Application & Data

Business Tools