We understand your dilemma and aim to cut through the clutter for you by shedding some light on three of the leading platforms that most often arise in our conversations. In ongoing dialogue with CIOs and data engineers, we’ve found three stand out from the rest and come up most often as finalists. These are Fivetran, Matillion and Keboola.
In this point-of-view, we deep dive into each to help you better understand the landscape and make the most informed decision possible for your organization. Let’s dive in!
A look at the platforms
Fivetran, Matillion and Keboola are all cloud-based data integration platforms that can help businesses extract, load, and transform data from various sources into a target system. In our advanced data strategy and integration work, these three are often short-listed as the desired tools of choice – tools that teams are already using to solve their biggest challenges or opportunities when it comes to data.
Let’s start with an overview of all three, before taking a deep dive into the component and capability set of each:
- Fivetran: Fivetran is a data connector Software as a Service (Saas) platform that helps move data from source systems to your warehouse. Its capabilities automate the movement of data out of an organization’s operational systems into their data warehouse. It’s an ideal tool for organizations that need an easy-to-use data integration platform with a wide range of pre-built connectors.
- Matillion: Matillion is an ETL/ELT software designed to load, transform, and sync data across the cloud – and offers a range of pre-built connectors. Data connectors are features of the platform, but not the only thing it does. With Matillion and a data warehouse (such as a Postgres database, Snowflake, etc.), you can manage your entire analytics operation through it. It is frequently chosen as a tool of choice for cloud migrations as it is highly scalable.
- Keboola: Keboola is a highly flexible data platform as a service that focuses on the entire data operations process. It allows users to build their own custom connectors and integrations using a variety of programming languages and tools. Keboola’s features are generally on par with Matillion, but they go a step further by providing the data warehouse component (which can be a large expense).
It’s important to think through what platform is best for your company and business needs. There are key differences between each that serve a different ambition, while each offers a slightly unique value proposition based on your goals, current data maturity and level of support that may be required. Also, each tool has a slight edge for different industry types, which can potentially serve as an advantage lever for you in your journey. In the end, it’s not about finding the ‘best’ tool, but really about determining the what is best for your organization’s unique situation and use cases.
Feature-set deep dive of each tool
When Fivetran covers all of an organization’s needs, it works beautifully. The platform requires minimal set-up and configuration, so it is easy for non-engineers to set up integrations with limited technical expertise. Fivetran also provides over 150 pre-built connectors to a wide range of data sources, including popular databases, cloud applications and APIs. This greatly reduces the need for custom integration work. Fivetran can be cost effective for early-stage organizations (although there are additional costs for third party tools). Finally, with Fivetran, customers benefit from a fully managed service with automated updates and infrastructure management.
However, as a large organization, users often find they lack a personal touch from Fivetran customer services. Unless you have a significant contract, it can be difficult to escalate issues. In the same way, it is unlikely that you will get support from Fivetran to build a new connector; so, unless you can utilize one of their pre-built tools, you may need to purchase an additional solution. And, while Fivetran offers some limited data transformation features, it is primarily focused on data extraction and loading – unlike Matillion and Keboola.
Fivetran can be a cost-effective solution for early-stage organizations, but add-on costs such as a data warehouse will be incurred. Fivetran is missing other important features of the data technology stack that are vital to leverage your data, such as a transformation framework, data warehouse, and orchestrator.
Suggested industries: Fivetran’s simplicity makes it a great choice for modern customer environments or “West Coast Startups” where off the shelf software is typically used. It’s wide range of pre-built connectors are perfectly suited to e-commerce businesses (i.e., connectors to Shopify, Magento and WooCommerce) marketing analytics (connecting to popular advertising platforms like Google Ads) and SaaS businesses (pre-built integration with apps like Zendesk, Salesforce and HubSpot).
Matillion offers a flexible licensing model with different pricing tiers based on needs, allowing customers to choose a plan that fits their business requirements. Matillion is also highly scalable, and the platform can handle large volumes of data and complex data transformation operations – along with a graphical UI which is more familiar to traditional ETL users. When it comes to cloud migrations, Matillion is frequently chosen as a tool of choice to support lift-and-shift migrations from traditional on-premises solutions.
Like Fivetran, Matillion customers will have to pay additional costs for a data warehouse. Additionally, we sometimes see customers struggle with poor git integration, which can lead to unsatisfactory ongoing maintenance and development on the platform for testing and breaking.
Matillion’s approach to data integration is serving low-code tool features. Their preference is to handle your data transformations using their toolset, rather than designing with a code-first approach. This may be advantageous, or could frustrate your engineers, depending on the organization and culture. Matillion’s UI and design have the feel of more traditional ETL tools such as Microsoft SSIS.
Although Matillion has fallen out of fashion over the past few years, it is still a company to watch. They have a long history and could leverage their experience and relationships to continue to push further into the space. We see this in their recent efforts, for example, with a new dbt integration, and a soon anticipated offering for a more SaaS like offering.
Suggested industries: Matillion is often used by more transitional businesses migrating from ETL/ SQL environments. With its advanced data transformation capabilities, Matillion is well suited for healthcare organizations, financial services, and manufacturing. For example, it can help organizations collect and analyze patient data from electronic health records, consolidate data from trading platforms and risk management systems, or integrate real-time data from systems across an entire supply chain.
Unlike Fivetran and Matillion, Keboola allows users to build custom connectors to suit their own business requirements; this makes it a popular choice for companies needing more flexibility. Keboola is the only platform on this list that does not require additional tools to support a data stack. It also offers a fantastic ecosystem for analytics, data engineering and data science, and supports a variety of scripting languages (SQL, Python, R), giving you the option to fully customize the data pipeline as needed. For this reason, Keboola is commonly used for businesses looking to build comprehensive analytics with strong cost management. While the implementation requires technical know-how, once set up it provides all-in-one solution for data management.
Keboola has more of a code-forward approach, which is in contrast with Matillion. Their platform comes with a very full suite of API’s, a Python library for building your own components, and a CLI tool where your organization may want to customize its workflows. Keboola also has development environments, change management and versioning built in. This is a useful feature for data teams that do not have a software engineering background where nowadays, change control is highly controlled.
Finally, while Keboola has been doing this for many years, there are not many SaaS providers that will have the option of providing the data compute and data warehouse. For many organizations, this makes it easy for teams to go from nothing to something, very quickly, within one contract.
On the flipside, Keboola employs its core engineering team in Europe; this means level 2 support may have time zone difficulties for companies outside of this region (although premium support with SLAs is available for an additional monthly price). Finally, while Keboola has built-in integrations, sometimes technical skills are needed to extract or push data into applications.
Suggested industries: Keboola is frequently chosen by early to mid-stage business for “all in one” benefits and total cost of ownership. With the ability to handle unstructured data and build custom integrations, Keboola is excellent for clients in media and entertainment (sourcing data from social media and advertising platforms), energy and utilities (consolidating information from smart meters, sensors, and other energy systems) and government organizations – enabling better decision making and policy implementation.
Your business goals ultimately determine the final solution
This is a complex space and selecting the right tool is essential to maximize the power of data within your organization and derive great value from it.
Our analysis and experience conclude that Fivetran is likely a stronger tool suited for businesses that need an easy-to-use platform with pre-built connectors. However, with Fivetran, you will probably need to string together 3-5 additional tools and vendors. Matillion might be an ideal play for companies in need of a graphical user interface for transformations. With Matillion, you will need to host it on your own servers (requiring setup and management), and you need to have a data warehouse technology. Finally, Keboola may be more ideal for companies that require greater flexibility in building custom integrations and advanced and complex data workflows. If you utilize Keboola, you don’t need to buy anything additional as they offer a data warehouse solution with it.
At Cuesta, we have rich experiences in this space working with organizations of all types, and we would be happy to help guide you along in your data journey and share more of our expertise.