Beautiful UI and UX to compete with any ETL, but at a cost!
October 06, 2020

Beautiful UI and UX to compete with any ETL, but at a cost!

Adel Helal | TrustRadius Reviewer
Score 3 out of 10
Vetted Review
Verified User
Review Source

Overall Satisfaction with Matillion

We still have an old-fashioned process in our monolith environment to extract-transform-load from a SQL server data store. Currently the monolith is the main product of the CM brand and is still being heavily maintained. The way we are moving forward is to "unlock" the data from its current storage in order to have better analytics using event streaming. For the time being, the only way we are able to extract large amounts of data from our relational SQL data store is using the ETL process through Matillion into our data warehouse solution in Redshift.
  • The easy-to-use GUI makes it easier for our team to pass on the knowledge and upskill engineers on our ETL processes.
  • The feature set is rich with many options to allow us to try different ways to transform our data without having to code.
  • Many different integration points allow us to plug straight into services like SQS to help us communicate with our own internal services.
  • Matillion does not scale well. It has a hard limit on the hardware / EC2 instances it can use. Most of the time that does not provide enough parallel processing for the millions of records we want to transform.
  • It is expensive considering the infrastructure cost is added to Redshift costs, so the overall value for analytics is something we are constantly challenging.
  • Constant Java heap space errors, again this is because of hard limits on EC2 instance hosting.
  • Considering our specific use case is to provide analytics "for free" to our existing customers of all tiers, the added investment is hard to measure since this is an internal cost in order to keep up with our competitors. We are still trying to quantify customer retention directly related to the new analytics we provide because of what Matillion has helped us achieve.
Extremely easy to use, extremely well documented, very familiar workflow layout to other comparable ETL tools out there, very interactive drag-drop features...
From the moment we bought it on AWS marketplace to connecting to our monolith database it took a matter of hours. Deciding the Redshift table structure we needed for storage actually took longer than the actual setup time! That is a definite plus for anyone who wants to get up an running as soon as possible.
That is the one sticking point we have with Matillion. We have hit its hard limits and constantly face Java Heap space errors. Our recommendation from the support team was to spin up another new Matillion server which is a separate cost entirely, making it equivalent to buying two separate products off the shelf, rather than incorporating it into the one solution. One Matillion server is expensive enough, two is not feasible.
It is much easier to use in terms of GUI capabilities. The only reason we would use an ETL tool other than our own manually written SQL scripts, is to be able to allow other engineers to use it without having one domain expert stuck on the inner working of complex scripts. So the GUI is essential for that, otherwise we wouldn't bother considering an off the shelf product.
Essentially Matillion is a nice tool to build a bunch of Redshift queries with added benefits of having a range of integration features. Considering the annual cost is almost equivalent to a junior engineer, we could spend the time to manually write the Redshift queries we need to transform and load our data that can run on our own microservices, and scale them as we see fit so we aren't faced with the hard limits of EC2 instance sizes. We also wont have to worry about hitting Java heap space limits since we can focus more on the Redshift query/transforms optimizing.

Matillion Feature Ratings

Connect to traditional data sources
9
Connecto to Big Data and NoSQL
3
Simple transformations
9
Complex transformations
9
Business rules and workflow
8
Collaboration
9
Testing and debugging
3