What is your suggested SaaS analytics stack for measuring KPI's and actionable metrics?

byDan-ya Shwartz
Updated May 7, 2017

There are two different approaches to analytics - custom and off-the-counter. I’m a personal fan of custom analytics, so for me the main ingredients of a stable, strong analytics stack are:

  1. Data collectors and data sources across all domains - click stream, advertising data, usage data, operational feeds, CRM, etc.

  2. Data pipeline - also known as ETL - the process in which you move the data to the data warehouse

  3. Data warehouse- where all the data is stored ready to be analyzed

  4. Data visualization and BI - to help you analyze and gain insights from your data (preferably in a collaborative manner)

  5. Data application tools (optional) - when you actually want to use your data for something other than analytics

Here is my analytics stack:

Data collectors / data sources

  • Salesforce.com (product) for CRM , Delighted – Customer feedback with Net Promoter Score for NPS

  • Google AdWords, Facebook Ads for advertising

  • Javascript SDK for website tracking and user level data, Google Analytics (product) for aggregated view of the same data in a non-professional way

  • SERPs.com for SEO data

  • Postgres Databases is where we store our users activity

  • Marketo (product) and Mandrill (email delivery service) for communication with our prospects and customers

and many others…

Data pipeline / ETL

The role of a data pipeline is to help you get all your data in to one place. Breaking the silos is an important part of being able to reach deeper and meaningful insights, but it doesn’t end there. The data quality and integrity should be done as part of the ETL process too (unless you are one of the weird people who prefer data lakes, or should I say - data swamps :) )

I use Alooma as my data pipeline. I also work at Alooma. But to be honest - I only joined the company because I needed a proper pipeline to work with.

Here is Alooma in a nutshell:

Data warehouse

The concept of a data warehouse is pretty self-explanatory. I work with Amazon Redshift for 5 years now - but some of my best friends work with Google BigQuery and other flexible data warehouses.

Data visualization and BI

There are SOOOO many tools out there, but my personal favorite is Redash because it is:

  • Shareable & transparent - you can share a link to a visualization / analysis with the query that yielded the said result

  • Simple & straightforward - SQL is enough and all you need to know. No abstraction layers, just a good-old direct access to your data

  • Collaborative & comparable - a sort of github for analysts

  • The data is accessible via API, so you can build services based on your analysis

  • It connects to any data source (including import and export to Google Sheets)

  • It’s for amateurs and pro’s alike

Like what you read? Share on

Published at Quora. See Original Question here

Data AnalysisAnalytics

Further reading

What is Striim?
Alooma Team • Updated Jul 26, 2018
What is Alteryx?
Alooma Team • Updated Jan 1, 2018
How do I handle unstructured data?
Maytal Shamir • Updated Dec 27, 2017
What is a data lake in the context of big data?
Dan-ya Shwartz • Updated Jul 3, 2017