How (& Why) Retailers Can Do Big Data in the Cloud

Cloud computing changed everything for retailers. From brick-and-mortar enterprises to web-only storefronts, the cloud has had a major impact on business, technology and the bottom-line.

That’s why for most retailers today, the question is not usually “if” leveraging the cloud is a good idea. Now the more common questions are “how” to best leverage the cloud and “which” applications and processes to move first. Migrating analytics, big data processing and data warehousing to the cloud should be at the top of the list for retailers. If you think those sound boring, think again.

Adding cloud-based data processing capabilities gives retailers the agility to react to changing market conditions; drives deeper relationships with customers; and, often delivers major cost-savings and operational efficiencies. While moving some or all data processing to the public clouds (e.g., Amazon Web Services, Microsoft Azure, etc.) used to be a major technical challenge, new big data services have made projects easier and more cost-effective for enterprises.

Early security fears have subsided, as experts now say cloud infrastructure providers have more expertise and controls than ever, particularly compared to self-run data centers. With the potential upside, retailers cannot afford to wait on big data in the cloud.

Why Process Retail Big Data in Cloud?

One of the most compelling reasons to move processing to the cloud is shifting “data gravity.” While a lot of a retailer’s data used to be generated on-premises at offices or storefronts, significant volumes are now generated in the cloud – think digital, web, social, mobile, Internet of Things or software-as-a-service applications.

These new cloud sources are often “big data,” that is, with much higher volume, variety and velocity than typical CRM or POS data. With the gravity of data shifting from data centers to the cloud, it’s more efficient to move processing closer to the data, rather than move all that big data back to an on-premises data nter.

Strategically, cloud data processing is a boon to agility. New cloud data infrastructure can be added in minutes, thanks to automation, elasticity and cloud services. On the flip side, building new infrastructure in a datacenter is a major operation, requiring in-house expertise, hardware and months of integration and optimization work. The cloud speeds up project delivery and delivers faster business outcomes.

Using the cloud for big data helps retailers:

  • Improve agility and access to new data sources: With traditional infrastructures, it can take months to make a new data source available to analysts, such as mobile app logs or loyalty data used to understand customer behavior. That long “time to analytics” translates to delayed insight, and lost revenue or cost-savings opportunities. But with cloud services, retailers can provision new infrastructure, load data and set access permissions in minutes, so workers can get to their data immediately.
  • Share data with stakeholders, launch new products: Retailers have long tried to share data throughout their ecosystem with suppliers and manufacturers, but it’s typically a long, manual process. Companies can dramatically improve data sharing by putting new datasets in the cloud for easy, secure access by customers or partners. With the cloud, companies can provide new data and analytical capabilities, in some cases, even monetized behind new applications.
  • Combine data across sources for deeper insight: Analytics across different data sources often garners the most interesting insights. At Cazena, we recently worked with a retailer on a new cloud-based data mart combining data from ERP, loyalty, POS and HR systems. From this foundation, the retailer will be able to use a variety of analytic tools and methods to answer a long list of questions about customer segmentation, the impact of loyalty programs, workforce planning and efficiency, inventory forecasting and more. Since the new data mart is in the cloud, it can instantly scale in size or power as the company grows and adds advanced analytics.
  • Collect customer data from many sources: Customer behavior data is available from a huge variety of sources: every web click, every mobile app tap, social interactions, in-store actions (transactions, beacons, mobile check-ins, etc.) and third-party data services. Given the exponential scale, speed and growth of customer data, much of it outside enterprise firewalls, the cloud is the natural place to collect, combine and analyze many sources of insight.
  • Improve omnichannel experiences and marketing programs: Targeted interactions synchronized across channels rely on a variety of data sources, many of them in the cloud. Companies are using big data to go beyond the classic 360-degree view, seeking to understand not only each customer, but also who those customers know and other indirect purchase influencers.
  • Cut costs (potentially!): Using the cloud can significantly cut infrastructure costs, if correctly provisioned and managed. Cloud services take advantage of economies of scale, automation and centralized expertise of infrastructure providers. Other benefits are elasticity, no capacity planning, easier administration, increased security … and the list goes on. The cloud can bring new agility to old retail processes, too, like the vendor that is now able to cut delivery time of datasets to its partners from months to hours.
  • Focus people and resources on strategic analytics: Moving processing to the cloud can also greatly reduce database administration requirements. Resources can be redeployed to focus on analytics and innovation, rather than managing and supporting infrastructure. Related, many companies hope to shift spending from capital expenditures (CapEx) to operational (OpEx) cost models and “as a service” offerings. However, while cloud economics help significantly, it’s no guarantee of cost savings. Be forewarned that provisioning the cloud is a black art that requires specific expertise, internal or from providers.

How to Get Started with Big Data and the Cloud

There are good reasons why many retailers haven’t moved data processing to the cloud yet. Many have decades of legacy systems built up, including data processes that may have taken years to architect. It’s been challenging to connect these on-premises systems and processes with new data infrastructure in the public cloud. The cloud is also a different operating environment for enterprises, with new pricing models, security controls and optimization tactics. Often, retailers have piloted the cloud, but found it challenging to put into production. This is changing, with more enterprise services that make the cloud easier to consume.

Here’s are a few key steps to get started:

Understand or benchmark your existing Time to Analytics: An emerging metric for data-driven companies, Time to Analytics (TTA) measures the time between a) when a company gets access to a dataset to b) when the right workers (often analysts) have access to that data so they can use it to add value to the company. For companies with on-premises data infrastructure, this is often measured in months. Consider recent data projects and their approximate time to analytics – and set a goal to improve that.

Evaluate cloud services that accelerate deployments. For example, Big Data as a Service is a new cloud category, recognized in recent analyst reports by both Forrester and Gartner:

“Enter the concept of ‘big data as a service,’ where vendors are combining components of analytic platforms in the cloud with multiple processing engines, hybrid on-premises integration, and secure data movement. The use of such services can speed up the adoption of analytics in the cloud, address skills shortages within the enterprise, and make it easier to transition from, and integrate with, existing on-premises investments.”

—  Gartner, Inc. Cool Vendors in DBMS. April, 2016. 

As you evaluate options, consider how services will work with existing tools and systems, specific security and governance requirements and functions for operational processes such as data movement between source systems and cloud, SLA monitoring and support.

Determine which data projects will make the biggest impact, culturally, operationally and strategically. Every organization is different. For some, the cloud may enable fast delivery of an innovative new capability. For others, the cloud may streamline key operational actions. Or, analytical access to a new dataset may unlock innovative ideas from workers. Consider: Now that cloud data infrastructure is easy, available and cost-effective, which projects will reduce unnecessary costs, inspire your workforce and delight customers?

Prat Moghe is the Founder/CEO of Cazena

This article was originally published in 2016 and is frequently updated