Matillion: ELT vs ETL
May 18, 2019

Matillion: ELT vs ETL

Clark Huang | TrustRadius Reviewer
Score 9 out of 10
Vetted Review
Verified User

Overall Satisfaction with Matillion

We are using Matillion within the Analytics department. We had a core need to be able to move data from multiple single-tenant client database sources into a proper data warehouse (AWS Redshift) for analytical reporting.
  • Easy drag and drop logic/control functions.
  • Ability to script (in Python) when out of the box components are not enough.
  • ELT vs ETL allows for super fast transformations done directly in Redshift.
  • We have had issues with out-of-memory errors when Matillion is up and running for a long time. For this reason, we've implemented an automated monthly restart job which works around this issue.
  • We do a lot of "reverse ETL" processing. For certain use cases we need to run extracts out of the analytical data warehouse, massage the data, then move it back to our transactional databases for certain operational tasks. Although it is possible with certain components in Matillion, there could be more enhancements to those components to make life easier for some tasks.
  • Matillion was the key to be able to build on customer analytics feature
  • About a year ago we had to upgrade to a more expensive instance/license of Matillion due to performance issues. We wish that we had the ability to just run our Matillion on bigger EC2 boxes that have more memory capacity, as that was our primary constraint, not concurrency of multiple users using Matillion. They only have a few license levels and the ROI is not completely there for moving up to higher level licenses.
It was very easy to get started with Matillion. Their in-app documentation on the components, as well as more detailed external knowledge base, was good enough for us to build out our entire ELT pipeline with minimal outside help.
We were able to get up and running with initial ELTs within a couple of days.
As mentioned before, we have run into some memory leaks errors in the past when dealing with really large data sets, iterations, and transformations.
We evaluated Fivetran and Xplenty prior to choosing Matillion. Matillion was the only solution that easily satisfied our single-tenant to multi-tenant use cases.
If you have lots of data from multiple transactional data sources and schemas that you need to iterate through and combine into a data warehouse, then Matillion was the easiest solution we found when we were in our proof of concept phase 3 years ago. We had some performance issues initially, but Matillion support and account management were able to help us through those, and now it has been running very smoothly for the last year.

Matillion Feature Ratings

Connect to traditional data sources
Connecto to Big Data and NoSQL
Simple transformations
Complex transformations
Data model creation
Metadata management
Business rules and workflow
Testing and debugging
Integration with data quality tools
Not Rated
Integration with MDM tools
Not Rated