As part of data democratization and the transformation towards data-driven management, more and more companies are adopting solutions that support and strengthen this culture. Data is made accessible to a wide range of users for analysis and reporting in tools like Power BI. At the ETL level, through which we perform data integrations into a central repository – a data warehouse, there is also a fundamental change happening. Cloud tools like Keboola (link to the website) are emerging, which do not require programming or administration knowledge (installation, server) and thus provide data integration opportunities to a wide range of enthusiasts.

Keboola Self-Service Data Platform – Introduction

Keboola is a startup providing cloud services focused on data integration in a self-service mode. This means that this tool allows you to move data from point A to point B (ETL processes, pipelines).

Founders Pavel Doležal, Petr Šimečka, and Milan Veverka are on fire. At the end of 2023, Keboola received a massive investment of 700 million CZK (approximately 30 million USD) from the American fund – Viking Global Investors. It seems that Keboola has already completed an extensive development phase and is now likely trying to compete against its biggest competitor, Fivetran, in the field of Sales & Marketing. Fivetran is also a cloud platform working on a very similar principle. We’ll see how this battle for market share unfolds🔥 However, Keboola is well-funded, and the market for democratization of business intelligence tools is heating up. 1

Keboola Products and Features

Keboola prides itself on being a tool accessible to a wide range of users without programming knowledge. We will hopefully see that in a series of Keboola tutorials that I am preparing. One of the significant benefits for users is that it includes a large number of predefined connectors to data sources and destinations (more than 700). Developing data integrations is much faster and more efficient.

In addition to the ability to connect to data sources, Keboola supports various useful features that allow efficient and secure data work in the environment of data warehouses and datamarts. For example:

  • 700+ connectors for various sources and destinations – databases, storage, ecosystems (AWS, Google, Azure, etc.)
  • Native support for dbt – a popular tool for data integration and orchestration
  • CDC (Change Data Capture) – identification of increments
  • Data catalogs – records of data to help teams better understand the data
  • Support for Data Science
  • Data lineage – analysis of data integration dependencies
  • Keboola as code – the ability to control Keboola pipelines from the command line/terminal via API
  • and more

Keboola Freemium and Pricing

Keboola offers a Freemium model. This means that the basic tier is free and is limited in terms of features and the number of minutes you can use per calendar month. 2

So, what do we get in the Freemium version?

  • 120 minutes of computational power for the 1st month and an additional 60 minutes each following month
  • Unlimited number of ETL/ELT workflows
  • 200+ data connectors
  • Support for transformations using (SQL, Python, R)
  • Analytical workspaces (SQL, Python, R)
  • Backend environment of Snowflake
  • Orchestration capabilities
  • 250 GB data storage
  • 1 Keboola connection project

After exhausting the computational power, you can purchase additional machine time for 0.14 USD/minute, or around 8.4 USD/hour.

Cost estimate: If we were to use Keboola purely as an extraction tool (source-destination 1:1), with “expensive” data transformations done elsewhere (e.g., locally), what could be the daily consumption?

Let’s say we have, for extraction, 200 source tables, each running for an average of 30 seconds, as we have incremental data implemented. That’s 100 minutes, or 1.5 hours. That would be 12.6 USD per batch. So, we’re looking at approximately 380 USD per month. That’s 4.550 USD per year + cost of data failures and recalculations (20%) so 5.500 USD. This amount is without considering transformations and server costs. Not bad.

If we use Keboola for everything, we will logically pay even more. “Everything” includes:

  • Data extraction from source systems
  • Identification of increments (if not available)
  • Historization (stage, trusted layer)
  • Calculation of the semantic layer
  • Calculation of datasets
  • Streaming

There, we could reach different amounts – multiples.

! Final amount, of course, depends on the specific conditions of the given company and IT/company environment (systems, complexity, data requirements, business niche)

Conclusion and First Impressions

Recently, I did a detailed review of the open-source tool Mage.ai, which is a fantastic tool for data developers that require some Python and administration knowledge. Keboola targets a different segment – it allows almost anyone to integrate data (ETL democratization).

The advantage lies in this, but also in the economic risk. Data work requires knowledge – approaches, ETL architecture, security, SQL for efficient transformations, and powerful scripts. Without these skills, a company may eventually face pricing pressure (cost of inefficiency). So, these tools need to be operated by competent individuals.

I plan to thoroughly explore Keboola and prepare a series of articles, so stay tuned.

5/5 - (1 vote)

Použité zdroje
  1. Keboola, Keboola vs Fivetran: Reduce operational costs without slowing growth [online]. [accessed 2024-02-02]. Available from WWW: https://www.keboola.com/lp/fivetran-vs-keboola
  2. Keboola, Keboola Pricing – Build end-to-end data pipelines in one platform [online]. [accessed 2024-02-02]. Available from WWW: https://www.keboola.com/pricing

Ing. Jan Zedníček - Data Engineer & Controlling

My name is Jan Zedníček and I have been working as a freelancer for many companies for more than 10 years. I used to work as a financial controller, analyst and manager at many different companies in field of banking and manufacturing. When I am not at work, I like playing volleyball, chess, doing a workout in the gym.

🔥 If you found this article helpful, please share it or mention me on your website

Leave a Reply

Your email address will not be published. Required fields are marked *