2#2 - The Business Value of Data (Eng)

MetaDAMA - Data Management in the Nordics

“The closer you are to the business, the great the chance to make an impact with data!”

In this episode I interviewed Marti Colominas, VP Head of Data & Insight at reMarkable. When we had our chat this summer, Marti was still working as Head of Data for Kahoot!.

Marti combines business with data and works on a daily basis for value creation and balance on the crossroads between business and tech. Marti has experience from big corp but was looking for that high pace and ever-changing environment of a startup.

Here are my key takeaways:

The Startup / scaleup setting

• In startup and scaleup roles are not that defined. There is a huge amount of flexibility.
• You can react quickly and explore new options without a lot of bureaucracy.
• It is very dynamic, and data is used by everyone to base their decisions on.
• On the other hand, datasets are not as stable and with less quality.
• There is a significantly shorter distance between C-level executives and Operations.

The Business Value of Data

• You need to find balance in delivering fast (time to insight, speed and accuracy).
• Speed is there no matter what, you need to ensure the right level of accuracy at that speed.
• If you want to have impact, you need good data quality. If not, the numbers will not be trusted or used.
• The goal is data that is trustworthy and easy to use.
• Long term commitment to Data Quality and Data Governance, whilst speedy day-to day operations with little time to insight, have been key to success.
• The role of CDO is to ensure that be business derives value from Data Science and Analytics, that counts for a Startup as much as for a large enterprise.
• The combination of business and engineering becomes more and more important.
• The data stack is moving towards speed and scalability, which makes it easier to handle large volumes of data.
• The key innovation will happen on automated data quality, self-serve analytics, even API-contracts for click-steam data, as well as tracking and lineage.
• Data is not always the goal. It can be a means to create value.
• For Kahoot! And reMarkable, data is used to make a better product and to improve the user experience, not to monetize or even mine that data.
• User should see and feel enhancements in the product through their provided data right away.
• To show the business value of Data Management you have to argue with "What if…?" What if we don't do it? You need to show the consequences of not acting to the C-level executives.
• Data quality problems grow exponentially over time. If you do not act, data-driven decision making will eventually be replaced by gut feeling.
• Not having control and metrics in place is like going to the casino: You can win once or twice3, but in the long run you will lose.
• Data products should have an assigned value - so Data as a product can help us argue for the business value of data.
• Step 1 is to treat data as an API contract.
• Self-service is dependent on a good structure and governance at the data producer side.
• Self-service can cover 60-70%, not everything can be self-service.
• A dataset tells the story of what happened in the business. You need business context to understand it, you need to phrase a business question into a data question, you need to know how to manipulate the data correctly.