In recent years, the data domain has gone through significant change, with the Modern Data Stack (MDS) playing a central role. What once stood for a major step forward in data management and analytics has gradually lost its status as a goal in itself. The focus has moved toward something more practical: data activation.
As this shift becomes more visible across the data community, questions arise about whether the term “modern” still fits a setup that is now widely adopted. A growing number of practitioners now refer to this setup as an “Analytics Stack”, a term that reflects a stronger emphasis on turning data into action.
The evolution of the Modern Data Stack
The Modern Data Stack emerged as organisations moved away from traditional, monolithic systems toward flexible, cloud-based alternatives. This transition was driven by the demand for scale, speed, and easier access to a wide range of data sources. The MDS answered this need by combining specialised cloud tools that cover the full data lifecycle, from ingestion and storage to analysis and visualisation.
Cloud-based architecture and its limits
The move to cloud solutions marked an important step forward. Teams gained more flexibility and could respond faster to changing analytics needs. At the same time, early MDS implementations still relied heavily on services from AWS, Microsoft Azure, and Google Cloud Platform. Building and maintaining these environments required deep technical expertise, which kept teams dependent on data engineers.
That dependency has decreased with the rise of self-service, cloud-based SaaS tools for ingestion, transformation, storage, and data delivery. These tools lower technical barriers and allow a broader group of users to work directly with data.
From Modern Data Stack to Analytics Stack
The shift toward an Analytics Stack goes beyond terminology. It reflects a broader role for data architecture: supporting dashboards, analyses, and activation tools that feed directly into business operations. The emphasis moves away from storing and processing data toward using data in daily decision-making and operational workflows.
This development aligns with a wider industry move toward activation-focused analytics, where insights inform tactics, support strategy, and improve operational performance.
As a result, SaaS tools play a larger role in data processing. Their use simplifies implementations and reduces complexity, allowing teams to spend more time working with data rather than maintaining infrastructure. Tool choice remains critical, both for the effectiveness of the analytics stack and for user adoption, learning curves, and long-term value.
Our preferred tools
At i-spark, we welcome the move toward activation-focused analytics, as it delivers tangible value for our clients in eCommerce and content-driven organisations. We have been delivering actionable data products for years, using advanced analytics tools that support this approach.
The tools we most often include in the composable data platforms we deliver combine strong functionality, scalability, and proven reliability. Each tool is selected based on experience, flexibility across use cases, and how well it fits within the wider ecosystem.
- Databricks provides a unified platform built on lakehouse principles, with governance, collaborative engineering, and data science capabilities. In combination with tools such as dbt Cloud, it supports advanced analytics and machine learning use cases.
- Snowflake is a cloud-based data warehouse designed for scale and performance. Its architecture supports elastic compute, efficient querying, and straightforward data sharing.
- Google BigQuery offers a serverless data warehouse optimised for speed and cost control. Its built-in machine learning features support analytical work directly within the warehouse.
- Fivetran automates data ingestion into platforms such as Databricks and Snowflake, keeping datasets up to date and reducing development effort.
- Dataddo provides no-code data integration, making data from SaaS tools and databases readily available for analysis.
- Hightouch connects data warehouses with operational systems, supporting both event collection and syncing processed data back into business tools.
- dbt Cloud acts as the transformation layer, turning raw data into analytics-ready models. It allows analysts and analytics engineers to prepare data independently while maintaining consistency and reliability.
- Looker supports advanced dashboards and reporting, allowing users to explore and interpret data directly from warehouses such as Snowflake or Databricks.
- Looker Studio focuses on accessibility and speed, with strong integration across Google’s ecosystem and simple data blending across sources.
- Klipfolio adds flexibility in dashboard and report creation, connecting to a wide range of data sources and supporting detailed performance views without heavy technical requirements.
An Analytics Stack never includes all of these tools. Instead, it is a carefully selected combination based on functional needs, technical constraints, and available budgets.
How we use the terms Analytics Stack and Data Platform
Our work with actionable data began well before the Modern Data Stack became a common term. Over time, we used various labels, such as data stack, data platform, and composable platform, to describe similar setups.
Today, we use “Analytics Stack” to describe the full set of tools that support data work from ingestion through activation, with a clear emphasis on actionable outcomes. The data platform itself refers to core systems such as Databricks and Snowflake, where scalable storage meets data and analytics engineering capabilities.
Looking ahead
The move from Modern Data Stack to Analytics Stack highlights a broader shift in how organisations use data. Analytics is now closely tied to business decisions and operational processes, calling for flexible, insight-focused architectures.
By selecting tools that support current goals while remaining adaptable over time, organisations can build data infrastructures that evolve alongside their needs. Terminology may continue to change, but one point remains constant: organisations that put data to work effectively gain a meaningful advantage as they grow.
