The Evolution of Data as a Service
The origins of data as a service (DaaS) can be traced back to the early 2000s with the emergence of web APIs allowing programmatic access to data. Platforms like Bloomberg, Thomson Reuters, and others began exposing financial data feeds via APIs targeted at developers.
As cloud infrastructure matured over the last decade, the DaaS model started gaining broader traction. Specialized data providers began offering curated datasets from sources like surveys, IoT devices, public records etc. on a subscription basis. Leading examples include financial data platform QUODD and pricing data provider Pricefx.
However, the current incarnation of DaaS powered by industrialized AI, big data and cloud represents an evolved paradigm. DaaS platforms now integrate disparate enterprise data, apply intelligence to enrich it, and enable widespread self-service access. As a result, they serve diverse analytics use cases from financial benchmarking to IoT analytics to knowledge mining.
The ability to extract and share value from data enterprise-wide makes DaaS central to data-driven digital transformation. A McKinsey survey found that 63% of organizations are piloting or planning to adopt DaaS in the next year underscoring its strategic importance.
DaaS Architecture and Key Components
Modern DaaS platforms leverage cloud infrastructure to offer data access rapidly at scale. While architecture details vary across providers, some key components include:
Ingestion Layer: Provides connectivity to 1000+ data sources from relational and NoSQL databases, files, streaming feeds to SaaS applications via pre-built connectors. Handles translating source data into standardized format.
Preparation Layer: Automates processing tasks like structuring unstructured data, cleansing to fix inconsistencies, enriching with external context, merging disparate datasets. Leverages machine learning techniques.
Storage Layer: Instead of a single data warehouse, uses combination of cloud data lake for cost efficiency and speed paired with purpose-built μServices for analytical workloads.
Consumption Layer: Enables self-service data discovery, access and insights generation through intuitive search, SQL, dashboarding, augmented analytics powered by AI and ML.
Detailed DaaS reference architecture (Image credit: Azure Architecture Center)
Cutting-Edge DaaS Use Cases
Data as a service empowers innovative use cases by making robust data readily available and easy to activate. Some leading edge examples include:
MLOps
DaaS is invaluable for MLOps use cases focused on continuous development, deployment and monitoring of machine learning apps at scale. DaaS platforms like Palantir‘s Foundry reduce the complexity in feeding ML training pipelines with quality datasets. And they simplify data lineage tracking critical for model governance.
Smart Products
Technology brands like Peloton, Nokia and Honeywell are creating smart consumer and industrial IoT products powered by data and AI. Embedded DaaS capabilities allow ingesting, analyzing and applying insights from sensor feeds and contextual data closer to devices. This drives personalized experiences while preserving privacy.
Inter-Organization Data Sharing
Companies often hesitate sharing data externally due to risk and lack of trust. Platforms like Ocean Protocol allow enterprises to access, trade and monetize datasets securely while maintaining control via blockchain-based democratic governance frameworks. Such decentralized data marketplaces expand collective intelligence.
DaaS Market Landscape and Segments
Gartner estimates the burgeoning DaaS market to clock a staggering 33% CAGR from 2020 to 2024, reflecting growing enterprise demand. From a size of $2.08 billion in 2020, DaaS spending is projected to reach $12.2 billion by 2024.
External DaaS
Exposing data externally represents a $5.3 billion market opportunity including:
Data Marketplaces: Snowflake Data Exchange, Dawex
Data Monetization: QUODD, Eagle Alpha, FactSet
Internal DaaS
Internal centralized data access and analytics represents the bigger $6.9 billion market covering:
Self-Service Analytics: Dremio, Kylo, Trifacta Wrangler
Data Fabric: Denodo, Tableau, Talend
Data Ops Tools: Preset, Domino Data Lab, Allegro.ai
Best Practices for DaaS Adoption
Organizations embarking on a DaaS initiative must holistically address people, process and technology elements:
1. Pilot Focused Business Capabilities: Work jointly with business stakeholders and start small. Launch DaaS access focused on high visibility analytics use cases.
2. Promote Data Literacy and Culture: Gradually ramp data democratization as skills mature. Incentivize desired data-driven behaviors aligning KPIs and performance management.
3. Modern Data Architecture: Evolve underlying data infrastructure applying leading practices around DataOps, data fabrics and knowledge graphs to maximize DaaS impact.
4. Change Management: Help users transition effectively via training workshops and online resources. Monitor platform experience soliciting user feedback to guide enhancements.
The Future with DaaS
Already 78% of enterprises plan to shift more workloads from data warehouses and data marts to DaaS over the next year reports TechTarget. And Gartner notes that by 2025, over 50% of analytics capabilities will leverage big data DaaS solutions up from under 10% today.
As DaaS capabilities mature in lockstep with cloud and analytics democratization trends, data is poised to serve ever more users enabling smarter decisions. The quest for embedding intelligence in processes, interactions and products will fuel next-gen DaaS growth with CDOs playing a pivotal role. Ultimately, DaaS allows “data refineries” transforming raw information into value and activation enterprise-wide.
