At A Glance
Many companies lack a strategy for getting maximum value from enterprise data.
Even among those that do have a data strategy, success requires careful planning.
Successful strategy execution requires scalable data analytics pipelines.
Using the modern cloud as a platform for those pipelines makes it much easier for companies to gain value from data.
Unfortunately, even when such strategies are in place they are at high risk for failure, whether due to lack of clear business goals, technology vision, infrastructure, available talent with the right skill sets, or a combination of the above.
For companies with the foresight to prioritize capturing value from data, a modern cloud platform can offer an escape from many of these risks and restrictions. For those with the ambition, a committed focus on “value-from-data” strategy is the heart of end-to-end digital transformation.
In this article, you’ll learn some of the considerations involved in a value-from-data strategy. Then, we’ll explain how a modern cloud platform helps address them.
Regardless of your business or industry, you have a lot of data, and your data volume is probably growing quickly.
This data — a potentially precious source for uncovering business insights — is usually locked into disparate silos, where it sits inertly as an untapped resource. Although many companies have progressed to on-premises data warehousing, those data warehouses are designed to make only a small structured slice of data available for analysis — and then only by a relatively small group of people, for a limited range of analyses, and in a batch-oriented way.
In recent years, the Apache Hadoop ecosystem has emerged to expand access to data with respect to doing large-scale data analytics relatively cost-effectively. However, the complexities and talent gap associated with on-premises deployments are well documented, so the obstacles to meeting value-from-data goals are not necessarily less intimidating.
The following are some important questions to answer when considering a value-from-data strategy, regardless of the platform you choose:
- What are the business goals for my data? Is it to improve business decision-making, enhance product quality, build consumer loyalty, or something else?
- Where does my data come from (transactions, server logs, third parties, devices/IoT, social media), and where does it reside?
- How does that operational data arrive (in batches, in near-real-time streams)?
- For real-time streams, how fast must your system include new data in reports? Seconds, minutes, hours?
- How much of that data can I afford to process and store in a data warehouse or other analytics platform, and how does it get there?
- How will I manage spikes in data volume, and spikes in queries?
- Is there a culture present that encourages data-driven decision-making across the organization (not just among analysts and data scientists)? Who should have access to the analytics platform, and what tools and skills do they need?
- Is doing machine learning for predictive analytics (e.g., for large-scale fraud detection) an aspirational goal?
Answering these questions will help you understand your restrictions and requirements. To summarize next steps broadly, the implementation phase will include the development of scalable data analytics “pipelines” that extend from multiple data sources all the way to the appropriate decision-makers.
In the on-prem world, that pipeline often looks more like tangled spaghetti, introducing the complexity that makes a value-from-data strategy such a challenge to execute successfully, and the development timeline can be quite long. It also commonly includes choke points where, whether for reasons of deficient architecture or cost, data doesn’t flow freely (e.g., can’t be stored, moved, or queried affordably at scale or in real time).
There is a new alternative, however, offered by advances in cloud computing.
Why the cloud?
Implementing these pipelines on a modern cloud platform frees you to focus on your business opportunities, and far less on infrastructure, resources, and operations (including security). And such a platform makes it easier to add new functionality, such as machine-learning models for doing predictive analytics, to the roadmap. Specifically:
- A modern cloud platform hides the associated complexity in this process, from the provisioning of compute and storage on demand, to the optimal processing and formatting of data (whether it arrives in batches or in fast-moving streams) for queries. Many of the integrations and operations are done for you in the form of managed services.
- It contributes to cost control in a similarly elastic way; you pay only for the resources you use.
- It offers more processing power, network bandwidth, affordable storage capacity, and security infrastructure than most companies could afford to build, which dissolves many of the familiar technical barriers to strategy execution. This purpose-built infrastructure also facilitates the abstractions and automation that support the managed services described above.
- It separates computing resources from storage resources, which offers much more flexibility than traditional, tightly coupled architectures.
- It encourages a culture of “citizen data science” at the end of the pipeline because more people — not just trained data scientists — have access to data and the powerful tools they need to uncover and share new insights about the business.
For companies aspiring to achieve end-to-end digital transformation on the modern cloud, using data and analytics to drive agility and insight in business decisions is an important goal. It’s also important to have the flexibility to accommodate data volume spikes, new data sources, “fast” (streaming) data, and new users as needed. This way, businesses can address rapidly changing requirements, such as new applications and markets, without the need to scramble for resources. Through long and painful experience, many companies have learned that this agility and flexibility is extremely difficult to achieve in the on-prem world, even with data center virtualization adoption having achieved maturity.
In the modern cloud, these kinds of goals are much easier to meet.
Many companies lack a strategy for getting maximum value from enterprise data. Even if your company has such a strategy, it can be difficult to execute within the limits and constraints imposed by on-prem systems. In contrast, implementing your value-from-data strategy in the cloud offers you more agile infrastructure and a wider set of tools, from data warehouse-based analytics to machine learning.
Whether you are creating a new strategy or revamping one that hasn’t delivered sufficient results, implementation on a modern cloud platform can deliver scale, speed, and agility that increase your opportunity for success.