Analytics-as-a-Service: Best Practices for Scalability

3AI April 27, 2021

Analytics-as-a-service offerings are based on broad categories and attract a range of players with different backgrounds. The urgency to transform into a digital business and to compete more effectively in the global market is forcing buyers to become more information-driven. However, there are still many challenges, including the lack of skills, a solid information management infrastructure, and an overall willingness to implement a Analytics strategy and Analytics initiatives companywide. Although many enterprises are still using Analytics at a tactical rather than a strategic level, market dynamics and new offerings from Analytics service providers are allowing buyers to explore new engagement models for their Analytics initiatives

Major segments include data as a service (packaged information), business analytics platform as a service (baPaaS) (infrastructure), managed business analytics (BA) (professional services) and software as a service (SaaS). Business leaders who influence analytics and data science purchasing decisions, rather than just focusing on the technology, care about solutions and approaches that focus on how analytics can add competitive advantage. As a consequence, analytics-as-a-service offerings that integrate data analysis with a technology platform appeal to these buyers. With the proliferation of hybrid cloud and on-premises applications, consuming analytics based on where the data sits can be more cost-effective. Nonetheless, much of the focus while buying analytics combines criteria for agility, speed, and time-to-market benefits.

Data As A Service (Daas)

Data as a service is a style of information architecture characterized by an organization’s information assets being made accessible on demand via a standard connectivity protocol, wherever they are located and on whichever platform they are stored. DaaS is equally applicable for internal and external data accessibility.

Data providers already offer customers access to their data via DaaS (as opposed to via downloadable datasets). Leading IT departments are migrating toward DaaS and away from heterogeneous application-specific and database-specific queries. Drivers behind this trend include: the need to simplify and standardize data access for an ever-increasing array of data sources; the Internet of Things; business users requiring a self-service means to access data; technology moving toward NoSQL data structures (such as Hadoop) for hosting structured and unstructured data; the emergence of the logical data warehouse; and information ecosystem borders breaking down among trusted business partners. DaaS is a general architecture behind the development of common data layers. Most cloud providers offer a degree of DaaS-style data access, rather than requiring customers to be knowledgeable about underlying data structures and protocols. As the benefits of open and self-syndicated data take hold, protocols for externally accessing these datasets or streams typically take advantage of the simplicity of DaaS.

DaaS architectures can offer benefits to developers who may otherwise have to learn specialized database calls and protocols. This can enable new business applications to be developed more quickly, and new developers (or consultants) to get up to speed more quickly. DaaS architectures can be used to migrate complex hub-and-spoke data warehouses and data silos, or to obscure their complexity with a simplified interface — enabling more users to have access to more data, faster. DaaS architectures also make coordination among business units and among business partners more seamless — enabling more investment in business functionality rather than data integration.

Business analytics platform as a  Service (baPaaS)

Business analytics platform as a service delivers analytic capabilities and tools as a service. Solutions are often architected with integrated information management and BA stacks. These include databases, integration capabilities and BA tools — or solutions that include only BA tools (for example, reporting and dashboarding) — leveraging autonomous cloud-based or on-premises data repositories.

baPaaS is gaining functionalities, as well as credibility and traction among customers. More vendors now have cloud-based solutions available and deliver a broader set of capabilities, such as interactive visualization data discovery and predictive modeling. Solution delivery is becoming more flexible and can be made entirely through the cloud or, in some cases, through hybrid models, where the analytic tools are cloud-based but the data repository is located, managed and queried on-premises. This delivery model avoids having to transfer, update or process large data volumes, and it addresses concerns about the sensitivity of the information being uploaded.

The combination of cloud-based platforms (baPaaS) and the domain-specific information delivered through reports, dashboards and analytic models — all managed by cloud vendors — becomes BA software as a service (baSaaS).

baPaaS products offload server infrastructure and the administration/support burden while allowing IT and business teams more time to focus on analyzing the data that drives business performance.

Compared with on-premises deployments, customers using baPaaS should experience faster time to value, lower initial costs, and less need to maintain expensive skills to support hardware- and software-rich analytic platform.

Buyer Appeal

• Ability to source services across the entire stack, including deployment and management
• Single-point sourcing of entire stack
• Perceived cost savings compared with clients managing Analytics on their own

Buyer Evaluation Criteria

• Strength of individual components and the ability to source managed services from a single vendor
• SLAs and uptime benchmarks
• Strong data backup and archival services
• Flexible recovery testing policies
• Proven data privacy management controls
• A track record of successfully managing application failovers following disasters

Managed Analytics Services

Traditional reporting services have been increasingly outsourced through “reporting factories” for descriptive and diagnostic analytics. This has consisted of outsourcing Analytics platform management and report updates. In addition, service providers have been offering data warehouse management and development as part of adjacent outsourcing options. New services have emerged to accelerate use of predictive analytics and incorporate data science for analysis to marketing, risk and other operations leaders. This has consisted of offering an analytic platform bundled with data integration with a wrapper of analytical consulting services. These solutions accelerate the ability for operations leaders to bring advanced analytic capabilities to their business management with the introduction of several quantitative disciplines (such as statistics, machine learning, operations research and computational linguistics). For the first time, computer scientists, operations researchers, statisticians and others are all willing to unite behind the banner of “data science” — which is a very profound development, and managed BA services are a first step in creating an enterprise offering.

Benefits for Buyers

• Ability to outsource data mining and data science jobs
• Accelerate solution into operations leaders’ toolkit through opex
• Primary focus on analytical outputs rather than technology stack

Evaluating Criteria

• Relevant solution offering based on whether outsourcing technology or an operational effort
• Service levels based on business outputs rather than technology operations
• Flexibility in staffing expert services based on business demand

Cloud-Based SaaS for Analytics

Cloud-based SaaS is a growth market and also a transformational force. Cloud-based SaaS is generally trending up the growth trajectory of the adoption curve, and it is still generating demand in many application segments. Many enterprises are still exploring the possibilities of cloud Analytics, and while many remain skeptical, there has been increased interest in the as-a-service model among buyers in the past year. Most interest is in private or managed cloud, driven by business users who prefer a simpler engagement model and more efficient services. However, many enterprises lack the skills and overall strategy to integrate cloud services with their Analytics initiatives and other on-premises capabilities.

Benefits

• Strong potential business growth, leading to little patience with the “IT way” of deploying Analytics
• Pay-per-use consumption model that allows business users to pay from their operating expenditure (opex) budget
• No upfront capital requirements
• Scale on demand and the ability to use the infrastructure for building business cases
• Rapid upgrade cycle

Conclusion – Which Analytics Vendors stand to win?

For businesses, analytics are about one thing: getting better. Service providers that understand the ways in which customers want to improve and provide the path to make those adjustments are most likely to develop successful partnerships with customers. As infrastructure and data services fall into place, providers can start to move up the stack, collecting stored data throughout different applications – even unstructured data housed on individual computers and drives.

As more companies look into analytics, service providers that find different ways to drive value and deliver results to customers are most likely to stand out as ideal partners. The types of businesses looking for better ways to analyze their companies and become more efficient are likely to become more diverse. Every company needs better information and better access to critical insights that transform operations positively. Analytics vendors with a strong understanding of customer needs and a clearly defined value chain stand to capitalize on the trend.