One day we'll look back at how we ran our data stack and wonder: what were we thinking?
What are the things data teams do today that will look embarrassing in five years - the habits that will make us wince when we look back at how we worked in the mid-2020s?
We asked analytics engineers, heads of data and a few reformed dashboard hoarders. Here is what they said.
Bob Muglia, Entrepreneur & Former CEO of Snowflake
Allowing end users to write queries. That’s what agents are for!
Ruth Onyekwe, Data & Analytics Manager at Spaulding Ridge
Central data teams. Having one team manage all analytics needs for an entire business without domain expertise.
Assaf Levinson, Sr. Director of Product Analytics at Gong
Having to write code in order to leverage ML models for data analyses.
Ross Helenius, Dir. AI Transformation Eng. & Architecture at Mimecast
The work we put in to get value out of data. We’re about to go through a condensed equivalent of the Industrial Revolution in knowledge workers. The velocity of this change will displace a whole swath of tasks we do today and open doors we didn’t think were possible for putting our data to work.
Gio Granato, Sr. Director of Data, ML & AI at Checkr
Excel. And we’ll probably still be using it in 2045.
Lindsay Murphy, Head of Data at Hiive
The modern data stack. Many of the tools we use now will either become features of other tools or be replaced entirely by AI agents. So spending money on learning and integrating dozens of tools to build highly decoupled data stacks with 15-20 different tools will soon feel absurd (within the next few years, if I were a betting woman).
Eva Schreyer, Head of Data & Analytics at Neugelb Studios GmbH
Data silos. When data teams worked completely disconnected from business teams.
Tristan Handy, Founder & CEO of dbt Labs
Migrations. Data teams spend a shocking amount of time simply moving existing functionality onto new platforms.
Jenna Jordan, Sr. Consultant at Analytics8
The very idea of “data” teams. That there is a central team that specializes just in data, rather than domain-oriented information professionals that combine data and technical skills with domain knowledge to produce cohesive knowledge products.
Tiankai Feng, Author of Humanizing Data Strategy
Defining what a data product is. In the future, all data will be treated as products.
Mark Nelson, Venture Partner at Madrona, former President & CEO of Tableau
Sending users “pictures of their data” Sharing data in PowerPoint or PDF rather than having them interact with their data and answer new questions with it.
Richard Cotton, Senior Data Evangelist at DataCamp
Being called a data analyst. Data tooling and literacy keeps improving so anyone can do data analysis.
Nir Smilga, Data Viz Manager at monday.com
Cluttered dashboards. Future data teams will laugh at how we drowned our users in endless dashboards, instead of crafting self-adapting data products that deliver the few insights truly driving decisions and impact.