Matillion ETL Reviews

We primarily use Matillion ETL to effectively manage data's movement, ingestion, and transformation through pipelines. We have specific use cases that involve different types of data, but they all fall under the general bracket of data movement.

We've got customers that have used Matillion to stand up entire and data platforms. We have customers that use it on a daily basis to perform heavy data movements and pipelines. We have a whole load of different case studies for different customers for different technologies. For example, Optio used Matillion to actually build a full new data platform and those pipelines. 

The most valuable feature is the ability to put together pipelines that push down all of the logic into Snowflake. So none of the actual execution or none of the data you need to travel through. But we can do it in a GUI-based system, which is a lot easier to hand over to end-users. Most end-users, in my opinion, have an easier understanding of GUI-based tools than code-based tools. And Matillion just has a whole load of features that assist with that, they get integration allowing you to have the exact same as slow but apply it to different environments or push code changes through two different environments is really useful. And their ability to leverage different forms of iterators over variables control tables, etcetera, so that you can orchestrate a whole variety of things in one go instead of having to kind of train them up.

So the main thing I would like to see improved in Matillion are two things. Firstly, their ability to process concurrent workloads. Right now, the concurrency reaches a stalling point if too many things are added, and it gets stuck waiting for each one to finish. 

Secondly, Matillion needs an improvement in its backend integration and the way that it pushes things through. It is already good, but it could be cleaner. I will say that I think both of those issues are being addressed in the new platforms that are coming out.  Matillion Unlimited Scale is the new answer to concurrent workloads, and Matillion Data Productivity Cloud is their new software-as-a-service version of a Matillion ETL provider, including a deeper git integration. So my concerns are being addressed, but those are the two things that stand out to me the most right now.

I have been using Matillion ETL for four years. 

It is a stable solution. However, the stability depends on how you host it. If you choose to deploy it on a robust virtual machine with ample RAM and CPU, it will be more stable. On the other hand, if you put it on a smaller virtual machine, it might be less stable. 

From an architecture standpoint, Matillion is as stable as any other virtual machine. However, the software has some memory leakage issues that may affect its stability. I suggest restarting your Matillion instance regularly to clear out the memory. It's also a good practice to shut down the instance when not in use. We have some customers who keep the instance on for only 20 hours a day and have a process that automatically shuts it off and starts it again four hours later.

It is a scalable solution, but with the new versions coming out, they call it "unlimited scale." So, the latest version of Matillion is a lot more scalable. The current version, in my opinion, is a bit more limited because it's on a virtual machine, and everything executes on that one virtual machine. If you have more users, you may need to deploy additional virtual machines, which can be quite cumbersome and require keeping them in sync. However, the new unlimited-scale approach means having one designer and one front-end. 

All the workloads are delegated out to Elastic containers, so it's very easy to log in to the configuration for the containers, change the number of containers that will be active, and scale out that way.

The maintenance part can be a bit of a challenge. Matillion has come quite far recently. It used to be that in order to perform an upgrade, you had to set up a completely new machine and migrate everything from the old one to the new one. That was a lot more of a headache. These days, they do support in-place upgrades. So, you can just click through a few app admin options, perform an update, and it will update the instance. However, if you go that route, you might still need to perform a full migration over the span of a year or so by setting up a new one before the full migration. It is because the more the underlying processes change, the more those in-place updates eventually fall behind, especially in terms of keeping the VM itself up to date. In-place updates only really update the Matillion application running on the VM and not the underlying libraries, etcetera. 

Over time, I have escalated quite a few things to the customer service and support team. They're good. They respond not just to our partners or us, but we've seen customers raise questions themselves, and they usually get a response from Matillion within a day. 

Matillion is always happy to jump on a Zoom call or something to try and resolve the issue. Also, they have an active ideas platform. So, if you go to the Matillion forum, there's an ideal area. They are quite good at taking the most uploaded ideas and implementing them into the tool going forward. I think there's a nice combination of the support dealing with existing issues and the developers trying to improve based on the community's input.

Positive

Only one person is required to deploy the solution. The process is relatively straightforward and takes around one and a half hours. 

The deployment model for Matillion ETL depends on the customer since we are a consultancy with several customers in different capacities. However, the majority of our customers use the cloud-hosted equivalent and host it in their own cloud environment. We don't have any customers who host it on-premises, although it is possible. Customers typically host it on their platform.

Additionally, we do have access to Matillion's own as-a-service beta, but it is only in a private preview environment. We only use it for general testing and experimentation.

For the deployment process in the current world, you would log in to Resilient and create an account on Matillion Hub. That is the process that allows you to get started with Matillion Hub. Matillion often helps you with this process if you are a Matillion partner. You go through Matillion Hub, put in your details, and then you can decide which platform you want to deploy your virtual machine to. You can choose between Azure, AWS, and GCP.

When you choose a platform, it will give you a template report bespoke to that environment. For example, if you're deploying to AWS, it will give you an ARM template. If you do it on Azure, it will provide you with a kind of shared instance template or a VM template. Then you just need to configure a few options, and it will deploy the instance to stand up for you. That's the current process.

In the new world, when Matillion launches its Data Productivity Cloud, things will change. It's delivered as a service, so I imagine the deployment will be much simpler. But my understanding is that it will still leverage your own custom zone containers to process the workloads. So the process will be very similar to what's currently used to deploy the virtual machines, but it will be used to deploy the elastic containers to which the workload will be pushed out.

I have seen ROI. For example, one of our customers in the UK has used Matillion quite extensively. They had a challenge where they were a group of different smaller companies, and they wanted all their engineers to work collaboratively on a single platform. That's where Snowflake and Matillion came in. They have one instance that's managing to serve all of these different sub-companies and sub-engineers, and they are easily recuperating the cost of that in order to provide data. But it is worth noting that they're not a profit-based company; they are a public health service, so it's more that they are saving money as opposed to making it.

The current pricing is based on consumption. So when you spin up a virtual machine, the size and type of the machine will determine the hosting cost by the provider, like Azure. 

Before considering the licensing cost of Matillion itself, you need to consider the cost of hosting the virtual machine, which is an additional cost. Matillion charges one credit per hour for each virtual CPU that the VM is using. So if you choose an 8-VCPU virtual machine and run it for 24 hours a day, the cost can add up quickly. The price per credit varies depending on your tier, but I think it's around $3.50 per credit for the top tier of Matillion and $2 per credit for the lowest tier. If you Google Matillion pricing, you can quickly find the dollar amounts. But you can manage these costs by shutting down your instance when you're not using it with an automated process.

In my opinion, none of the tools in this particular space right now are perfect, including Matillion. However, Matillion seems to be the best of the bunch. Its UI looks a bit dated at this point, but I find it reliable and relatively easy for users to get going with. 

There's not a big learning curve in Matillion. You can log in, and within about half an hour or an hour, you can know how the general platform works and how to get going. The main benefit I've seen that I haven't seen in some of the other tools is the ability to easily change which environment you're working with, such as swapping between dev, test, prod, or GitPrime and executing a flow in that location quite easily.

The main thing to consider before using Matillion is the specific reason you want to use a tool like it. If you need a combination of data ingestion and transformation, then Matillion can be a great option. 

Additionally, if you expect to load large volumes of data that will be used across multiple avenues, Matillion is a good choice. On the other hand, if you only need data ingestion and with more manageable volumes or ingesting streams, it might be better to look at a tool like Fivetran. However, they charge you based on the volume of rows or records you ingest. So, if you ingest large volumes, your costs can rapidly increase and overtake your maintenance costs.

Overall, I would rate Matillion ETL an eight out of ten. 

My primary use case involves handling standard ETL tasks. I work on processing both our company's data and third-party data sources. While I focus on these standard ETL tasks, my colleagues excel in more advanced pipeline work.

The most valuable feature of Matillion ETL is its user-friendly graphical interface.

As someone new to the data industry and with limited experience in ETL tools, I'm not familiar with other options. My background was as a university professor until about a year ago, so I'm still getting acquainted with this field. I found some of the more complex aspects of ETL challenging, but I grasped the concepts fairly quickly.

I have been using Matillion ETL for one year.

I would rate the stability a nine out of ten. It is quite stable.

The scalability of the solution is excellent. I would give it a nine out of ten. 
Matillion is our main ETL tool, and it is the one our consultants recommend. Currently, about 30 people at our company use it exclusively for all our ETL tasks.

My advice to new users would be to start by going through the tutorials and working with example data. These tutorials are quite helpful in getting you familiar with the tool. Additionally, try working with data that you understand well, such as data from your previous work or a familiar dataset. This way, you can focus on learning how to use the tool without having to figure out complex data problems simultaneously. Overall, I would rate Matillion ETL a nine out of ten. It works very well.

We use the solution for the ETL pipelines.

I believe the biggest advantage of using Matillion ETL is its speed of development. We don't have to deal with all the details; we can operate on the component level and do a lot with minimal effort. This is the biggest advantage; for example, we can look for particular components, configure them with some metadata, and simply run it. This simplicity and speed of development make it easy to do things quickly. Additionally, the graphical interface makes it easier to visualize something and find the details we are looking for, rather than going through a number of SQL codes and trying to find the issue. Therefore, the speed of development is the biggest advantage of Matillion.

The positive aspect of this solution is that it provides a graphical interface for jobs. Another advantage of Matillion is its extensibility; if something is missing, it can be easily adjusted or custom components can be written. What I really appreciate about Matillion is that it allows us to schedule jobs, so we can track and monitor their execution on a daily basis. I also like how the solution is organized in Matillion; everything is clear and visible, and we can easily access any details we are interested in. Additionally, there is a useful feature for reporting errors, so we don't have to worry about error handling within the jobs themselves. This feature is very convenient as it saves a lot of time and effort.

Matillion ETL has great Git integration that is perfect and convenient to use.

Unlike Snowflake which automatically takes care of upgrading to the latest version and includes additional features, with Matillion ETL we need to do this ourselves. Matillion upgrades the tool quite often, but we need to manually apply it in our environment. This manual process can be done in a few minutes, but it has room for improvement.

Recently, I needed to develop a component that runs queries on Athena, one of the AWS services. Matillion ETL does not have this functionality out of the box, so having an additional component to handle this would be quite convenient. The tool is quite flexible, and there is no source that cannot be easily integrated. The developers are constantly adding new functionality from release to release, responding to market needs. The only thing I was missing at some point was a component for Athena queries.

I have been using the solution for five years.

The solution has been available for a few years now. Initially, there were some issues, but the support was excellent. If something wasn't working, we could quickly get help to resolve the issue. After a few years of using the solution, it has become very stable. We don't have any problems with Marillion ETL; I haven't experienced any surprises. Matillion ETL is very reliable; whatever we develop works.

The solution is scalable due to its cloud environment. This is the beauty of the cloud; if we require a machine with more power, CPU, and memory, we can do it on the fly. We can simply go to the configuration and change the underlying machine, which requires a quick reboot. The new instance is then set up. This is more of a cloud-related feature than a Matillion ETL feature, but it is very easy to scale. If more power is needed, it can be done quickly and easily. It is also important to note that Matillion is usually connected to a database engine, such as Snowflake, AWS Redshift, Azure Synapse, or Databricks. Most of the processing happens on the database side. However, if there is external work such as loading data from S3 or moving data, there is some load on Matillion ETL. But the majority of the work is done on the database side because it is an ELT-like tool. The data is loaded onto the database and then the transformation happens in most cases. It is up to us how we develop it, but usually, the majority of the power is consumed on the database side.

I have had the opportunity to collaborate with the technical support team a few times and have found them to be extremely helpful. They are very responsive, knowledgeable, and adept at understanding the intricacies of our issue. In my experience, the cooperation has been perfect.

Positive

At the time I joined the company I now work for, two years ago, Matillion ETL had already been introduced. We wanted to switch to something nicer and decided to move to the cloud with AWS Redshift. We also wanted to use a graphical interface for development, rather than lots of code snippets, as it would be faster and easier to maintain. With the graphical approach, it is much easier to spot and go to the root of any issues.

The deployment time depends on the configuration, but with the cloud, everything goes smoothly and we can complete it within two hours.

I believe the cost is customer dependent, so depending on the instance we have and what we can have on-premise, we can have an installation in the cloud, or we can use the hub option. This is what Matillion ETL provides directly. A rough estimation of the cost is around 20,000 dollars a month, however, this is dependent on the machine used and how Matillion ETL is used. If there is only an hour or two of processing per day, it is better to use the hub approach as we only pay for the hours used. If there is ongoing processing, then an installation in the cloud is usually the better option. Ultimately, the cost is dependent on the individual case.

I give the solution a nine out of ten. This is one of the best solutions. Matillion ETL is closely integrated with the cloud environment, which is quite common, and thanks to that, we can take advantage of services available in AWS, GCP, and Azure. I have used a few solutions so far and this is one of the best. Everything works as expected, the tool is very intuitive, the monitoring is very well-developed, and the Git integration is great. From a developer or architect's point of view, it is quite intuitive and nice to use. Matillion ETL is one of the best.

We currently have 15 developers using the solution in our organization.

Depending on the specific case and requirements; when deciding, I would take into consideration what other options are available. From an end-user perspective, I really like Matillion ETL; it is comfortable to work with and easy to set up and maintain. The solution is not perfect, but there are no other similar solutions. The support is also very good and the integration with Git is quite nice, so it is quite flexible. Even if something is not supported out of the box, we can customize components or include Python code. It also depends on the amount of data to be processed, what kind of data it is, and the underlying database engine. There are cases where Matillion ETL makes sense, but there may be cases where it is not recommended. The solution is quite flexible, with the Snowflake dedicated version or Databricks, so we can decide which underlying data warehouse to use. If the company is GCP-related, they may not want to use Redshift, but there is an option for Bitquery. If developers prefer to use Spark, the Databricks version of Matillion ETL would be a nice option. I have been using the solution for a while and I feel comfortable with it. Compared to other tools like DataStage and Informatica, I can say a lot of good things about Matillion ETL, so it is quite convenient for us.

Bringing data from different sources onto our Snowflake data warehouse.

• It has helped us avoid a lot of red tape due to compliance. 
• It has helped us to get onto the cloud quickly.

• It is pretty easy to learn.
• No scripting is required.
• Being cloud native.

It is not an end-to-end platform for ETL. It brings in the data. To complete the pipeline, they might want to include some connectors which would put the data into different platforms. This would be helpful.

We are working with different platforms. Most of the connectors that we are looking for are included, but sometimes scripting is required. The good thing is we can reach out to support and have them create the custom connector for us. This can probably be improved.

We do not put much stress on it. We run it as a batch right now and spin it up on demand. So, we run it for three hours, then we shut it down. It is not real-time, and there is not a lot of streaming as of now, but it is running in the batch.

The size of our environment is not big. We started six months back. Right now, we only use one node, which is moving the data onto our data warehouse in Snowflake. This node is also very small at this time. 

Eventually, we will grow big and quickly, because we just had our three drugs approved. Therefore, a lot of data is going to come over. We will be moving this over to our data warehouse, which will need to increase significantly.

The technical support treats us well. They already have a support portal, and they are responsive, which helps.

We haven't used any of the on-premise solutions. What we used before was SQL Server Integration Service (SSIS), and when we were moving to the cloud, we wanted to use something that was cloud native (AWS).

The integration and configuration of this product in our AWS environment was excellent.

We are integrating it with SQL Server and the data coming in is from the on-premise file server. That's what we are connect with right now. Eventually, we will be pulling the data from Google Cloud and Twitter.

We are still in the process of moving the data over, so there are no data points as of now.

You probably don't even need to reach out to the company sometimes to purchase it. You can go to the AWS Marketplace. It's very easy to spin up, and the configuration is also easy. It spins up in your own AWS account. The only way you can get the product now is via the AWS Marketplace.

Their pricing is simple. They charge you by the hour. Whatever EC2 instance time that you have been running it for, that's what it's going to charge you for. The licensing is not based off of the per user or per server. They are going by the type of instance you spin up and how long you've been running it, so easy peasy.

The other products that we considered were SSIS and Informatica.

We chose Matillion because of the flexibility of the data and the company does not store data on its platform. As a life sciences company, which is HIPAA compliant, we didn't want to move the data over to the platform, so that's why we selected Matillion. It also has additional components that we did not have to pay for, as those features are included with the data. We can scale horizontally with it.

Give it a shot. See how easy it is to get started with the product, because the scripting which is required is minimal. Anybody who is familiar with the SQL Server platform and with SQL scripting can easily pick it up and run with it,

Overall, it is a really good product.

We use it for archiving and storing data.

"It has made our lives much easier," This is what my teammates who were using the other stuff before have said. Then, they moved it to Snowflake. Now, it is much easier and faster to use than before.

It is cost-effective. Based on our use case, it's efficient and cheap. It saves a lot of money and our upfront costs are less.

• Performance can be improved for efficiency, and it can be made faster. 
• Latency could be reduced. Sometimes, it takes longer to fetch data out of it. There are network issues because we experience a little lag.
• While the UI is good, it could be improved in its efficiency and made easier to use.
• It can be used by different consumers. So, I would recommend to the company to promote more, because people don't know much about it. If they promote more, they can sell it. They need more marketing.

It is pretty stable, but you can't blindly rely on it. On a scale of one to ten, I rate it around an eight.

It can scale to a great extent. It can handle the load that we are putting on it, which is about 5TBs.

The technical support was good at troubleshooting. The issue that was sent to them was resolved in a couple of days, so it was resolved quickly. 

The integration and configuration was pretty easy. It was completed in about two months.

It saves a lot of time.

I have heard from my manager and other higher ups, "This product is cheaper than other things on the market," and they have done the research.

Anybody can help to configure and train on it, learning how to use it. This will help speed up the process.

We use the AWS version only.

The solution is used for data ingestion from multiple sources to Snowflake and extract and load.

The solution's most valuable feature is the CDC (Change Data Capture) component.

Depending on the use case, the solution's pricing could be improved. Matillion ETL should include more enhanced capabilities for extracting data from the SAP systems.

I have been using Matillion ETL for more than one year.

Matillion ETL is a highly stable solution.

I rate the solution’s stability a nine out of ten.

Matillion ETL is a highly scalable solution because it is hosted on a cloud provider. So, that part is automatically taken care of. More than 100 users are using the solution in our organization.

I rate the solution’s scalability a nine out of ten.

The solution's technical support team is highly responsive.

Positive

Matillion ETL is a self-managed solution. Matillion deploys the tool. It's a SaaS offering that requires no extra deployment effort from our side. The solution can be deployed in a couple of minutes. Matillion's support team spun up the environment for us.

Since the solution is on the cloud and self-managed, your operational costs and admin efforts will decrease. The solution has saved around 15% to 20% of our operational costs.

The solution's pricing is not based on the licensing cost but on the running hours when the Matillion instance is up and running. Its pricing model is different from the traditional pricing models of Talend or Informatica. The solution was not expensive for my use case.

I rate the solution’s pricing an eight out of ten.

The solution's orchestration capabilities are pretty good. It does have an integration with DBT. I would recommend the solution to other users because it has a very clear roadmap with a good support system. The solution’s maintenance is done through an in-house team of three to four engineers.

Overall, I rate the solution an eight out of ten.

We are using Matillion ETL for extracting and integrating the data from different applications, such as SQL, and other data sources.

The most valuable feature of Matillion ETL is its ease of use. If you have had some experience with other solutions, such as Snowflake, the use of this solution will be simple.

The cost of the solution is high and could be reduced.

I have been using Matillion ETL for approximately six months.

Matillion ETL is a highly stable solution. We are using a stable version of the solution.

I rate the stability of Matillion ETL a nine out of ten.

The solution has limited scalability. However, for concurrent tasks, the solution has been scalable enough for our needs.

I rate the scalability of Matillion ETL an eight out of ten.

The support from the vendor could be better. The speed of the support could improve.

I rate the support of Matillion ETL a six out of ten.

Neutral

I have used Snowflake. However, I am beginning the using these types of tools.

The implementation of Matillion ETL is straightforward. The length of time it takes for the deployment of the solution typically can be completed in a few hours.

We did the deployment of the solution with some assistance.

The price of Matillion ETL is expensive.

For those who want a cloud-based analytics platform, I would recommend this solution.

I rate Matillion ETL an eight out of ten.

It is helping to make Makerbot a data-driven company.

The most valuable features are the components for SFDC, RDS, Marketo, Facebook, and Google AdWords; built-in verification; and scheduling, restarting & logging.

On a Redshift project, before Matillion was released, two people literally spent over one month using Sqoop to pull very wide data tables from Safesforce.com to Redshift. On a new project using Matillion, it took me 10 minutes to set up and begin importing data from Safesforce.com.

Built-in verification and sampling is a fabulous feature for ETL engineers. Anywhere along the transformation-pipeline, one can check, see and sample the data. This saves days & weeks of effort and leads to a far more agile project.

More frequent releases are needed, due to API changes from Google, Marketo, and Facebook. They frequently release upgrades to their API and consequently frequently deprecate the older version when only a few months old. The only way to use the Matillion components for these APIs successfully is for the Matillion release process to step up to the plate and have far more frequent "minor" API releases (as opposed to "major" product releases).

Even having these automated might not be a bad idea. Some customers willing to pay might open up a new revenue stream for "platinum" service, to take the headache out of this very valuable set of marketing components in Matillion.

I only encountered stability issues when accidentally performing EC2-intensive Python jobs (i.e., not Redshift-intensive SQL). These can kill the Matillion EC2 instance.

I have not encountered any scalability issues.

Customer Service:

Customer service is excellent.

Technical Support:

Technical support is above-and-beyond.

SnapLogic and Informatica: too slow; for MPP, they are just glorified and expensive Python schedulers.

Python scripts: high maintenance.

Initial setup was straightforward.

I implemented it myself.

We achieved ROI in <1 year.

The first two weeks are free; pay by the hour for smallest instance for next 2-3 months; after that, take out yearly discounted rate from AWS Marketplace for instance/devs in team.

We also evaluated SnapLogic, Informatica, Talend, and Hadoop.

The mindset of the traditional ETL tools is to off-load transformation to another server/DB. This totally misses the point of MPP and especially of Redshift. Load the data into Redshift early and then transform it inside Redshift ("ELT" not "ETL"). Matillion orchestrates the loading and transformation "pipelines", then gets out of the way whilst Redshift does what it is good at (i.e. the "grunt work").