A few days ago I was reading an article on cloud computing in retail. The article provided a number of retail specific use cases for cloud technologies. The article prompted a thought, so I dug into a few other articles on the subject. As I reflected on the examples provided in a half dozen or so articles it struck me that no one was really talking about the advantages or challenges of the cloud versus traditional architectures.
This led me to think about what are the true, primary cases where cloud (private, public, hybrid) offers a distinct advantage to retailers? When I say “distinct advantage”, I am referring to those cases where the cloud offers capabilities that a traditional architecture approach would struggle to emulate.
But before we jump into the use cases, let’s talk for a moment about cloud architectures. Cloud has many different flavors. In the public cloud space we have the traditional vendors providing software-as-a-service (SaaS), infrastructure-as-a-service (IaaS), and platform-as-a-service (PaaS) capabilities. Newly branded services such as database-as-a-service (DBaaS), storage-as-a-service (STaaS), disaster recovery-as-a-service (DRaaS), and others provide even finer grained service models. Regardless of the services provided, all public clouds are off-premise and are offered (typically) on a subscription basis.
Private clouds may be on-premise or off-premise, but are always dedicated to a single organization.
Hybrid clouds are essentially an orchestrated mix of private and public cloud environments leveraged to provide elastic capacity.
Regardless of the cloud model, it is essential that retailers architect for change. This means developing methods to encapsulate cloud functionality and reduce the cost (and time) of adding or switching cloud services. A common approach is the use of application program interfaces (API’s). Not only are API’s a lighter weight method of integration, a well constructed API can provide a way to open up cloud services to a broader audience inside the organization.
What are the inherent advantages of Cloud architectures? Some of the most sited advantages include lower upfront costs, more flexible cost models, geographic reach/coverage, robust disaster recovery, and frequent (automatic) software upgrades.
Let’s briefly discuss the upgrades and cost advantages. After spending over 23 years in retail, I can remember numerous (easily over 100) discussions in which a business unit or the finance organization wanted to delay an upgrade to reduce cost in the current fiscal year. Sometimes cost wasn’t even the issue. The business unit simply didn’t want any disruption during a busy time of the year and asked for upgrades to be delayed. A single year delay isn’t really a big deal. But what happens when it is then delayed the next year? And delayed again the year after? I recall numerous applications being five plus years behind their scheduled upgrade cycle. What happens when you upgrade a five-year old system? It is essentially a re-platform and is often very costly and disruptive. Not only do the delays increase the cost and disruption of the eventual upgrade, the delays also increase security risks as often times the operating system and/or supporting database systems can’t be upgraded as well.
Also, when you consider costs, it is important to consider ALL costs. Often overlooked when looking at the total cost of ownership are items such as operating system upgrades, hardware refreshes, supporting software upgrades (databases, browsers, etc.), and organization disruption. Organization disruption occurs when we incur downtime to install upgrades or have to conduct extensive business testing for new releases of the application. Cloud architectures and solutions insulate organizations from most of these costs and disruptions. Simply stated, a recurring subscription service is a much more predictable cost model.
We don’t discuss cloud disadvantages as often, but some do exist. For years CIO’s were concerned over performance and security. For established cloud services, many of these concerns have been alleviated. Performance and security are major components of any cloud service that has achieved large-scale use. However, lack of control and flexibility is still a concern to some. While cloud providers seek to be flexible, at some point they do force organizations to take upgrades. Also, many CIO’s are hesitant to place mission critical workloads on public infrastructure that may subject them to outages outside their sphere of influence.
As we consider the use cases that are the best fit in retail we have to consider all of the discussions on advantages and disadvantages. The five primary cases that come to mind are collaboration, e-commerce, analytics, inventory management, and supply chain visibility.
These are very broad classifications, and deserve more explanation. Let’s examine each one in more detail.
Collaboration activities are vital to any retailer. Whether it be collaborating inside the retail organization or collaborating with partners and suppliers, collaboration is an activity performed on a very frequent basis. Consider product designers in one location working with manufacturers or sourcing agents across the globe. Or retail supply chain groups working with suppliers on multiple continents to develop and refine product forecasts.
Cloud computing’s geographic diversity, virtual centralization of information and automatic upgrades work perfectly with collaboration use cases. Having computing instances closer to the user (geographic diversity) provides a better user experience and encourages greater use of the solution. However, since the cloud is distributed and virtualized, everyone is working on the same information, at the same time. And finally, since upgrades are automatic, there are no coordination activities that have to take place across all of participants. Anyone who has ever had to coordinate changes in EDI standards across a diverse vendor base will recognize this as a huge benefit.
On top of the idea of collaboration, let’s add the concept of supply chain visibility. Retail supply chains contain a variety of different players (retailer, manufacturers, consolidators, transportation providers, government (import/export), etc.). The ability to understand where product is inside the “glass pipeline” of the supply chain has always been a challenge. Cloud provides the ability to overcome some of the traditional challenges by addressing geographic diversity, performance, and standardization.
Obviously e-commerce has been around for quite some time. While it is entirely possible to operate e-commerce functions without the cloud, the cloud can dramatically improve many elements of an e-commerce deployment. The distributed nature of the cloud can be used to simplify content distribution and provide a better experience to all users accessing an e-commerce website. The cloud’s ability to operate seamlessly in “active-active” mode (geographically diverse servers running the same application at the same time) translates into more availability, less risk, and more sales.
Analytics are a hot topic in all industries and retail is no exception. Analytic applications often have a very volatile demand pattern. To explain this, let’s use an easy example. Anyone who has worked in retail knows Monday morning is a peak time for any analytics and reporting environment. On Monday morning analysts from all parts of a retailer (merchandising, marketing, store operations, supply chain, etc.) pull the results from the last week and weekend and conduct performance meetings to establish a plan for the coming week. However, the same analytics and reporting environments are in much less demand in the middle of the week. We used to joke that we had “build the church for Easter Sunday”, meaning the size of the computing environment was built for the largest single peak. This is a perfect scenario for cloud computing, which has the ability to expand and contract based on usage patterns. In addition, major cloud providers such as Amazon, Google, Microsoft, and IBM (Watson) have made significant investments in cloud analytics capabilities that would be very difficult for all but the largest retailers to emulate.
Having the right product at the right place for the customer is the lifeblood of any retailer. The process for managing inventory has always been a challenge due to the wide variety of markets and customers most retailers serve. Too many times we see some stores sell through products in a few days and suffer lost sales while other stores have abundant supply. Maximizing the investment in inventory and providing the highest level of customer service is the balancing act most retailers play every day. To do this, inventory can no longer be viewed by “channel” or “location”. Inventory must be truly fungible and systems must be intelligent enough to take advantage of inventory at any retailer node (location). Managing an accurate, centralized view of inventory is a critical enabler to providing the next level of customer service. In this case, the cloud provides high availability, high geographic performance, and seamless accessibility in a manner that would be difficult to achieve in any other architecture.
As we discussed before, retailers can opt to run any number of applications in a mixture of public, private, and hybrid cloud environments. However, for the five broad categories of applications we just discussed, the cloud offers unique and compelling capabilities that retailers need to consider.