BrainTrust Query: Cloud Computing – Can Retailers Rise to the Challenge?
By Bill Bittner, President, BWH Consulting
The National Retail Federation’s
ARTS (Association for Retail Technology Standards) division has been working
on a white paper on cloud computing for release in January 2010. The goal
is to give both business and technical people in retail several perspectives
of how this new technology model will affect them. The potential benefits
of cloud computing are significant. Technology delivery costs can be reduced
10 to 40 fold as centralized computing resources delivered over the internet
replace proprietary implementations. But there are also challenges to achieving
these results.
The ARTS paper describes two major types of cloud environments.
Technical clouds are strictly infrastructure and system software services
delivered over the internet. By sharing these resources, the retailer avoids
the internal personnel and training costs for maintaining their technical
environment and frees up the office space. The capabilities of these clouds
are limited by physics and the state of the art in technology at a particular
point in time. Vendors supplying technical clouds are thus forced into
a standard that is defined for them.
But as great as the technology benefits
may seem, the even bigger opportunity is the benefit to be achieved from
clouds that offer Software as a Service (SaaS). These clouds include the
technical environment and also offer the actual applications. At this point
SaaS clouds have generally addressed “commodity
applications,” those
applications with wide acceptance that require little integration with
existing applications or business processes. Companies can use Google applications
for email, word processing, spreadsheets, etc. and avoid needing to maintain
application software on individual PCs. If they don’t want to use Google,
there are cloud vendors supporting email servers from a variety of software
vendors.
The final piece to be covered by cloud computing may have the
most benefit but it will also be the most difficult to achieve. This is
where cloud vendors begin to offer retailers specific software services
that directly address retail operations. These include price management
applications that recommend and execute price changes; distribution applications
that plan, monitor, and control the flow of inventory, etc. The challenge
is that without the same constraints that standardize technology clouds,
the SaaS clouds may provide custom solutions that can only be used by a
small number of retailers. The whole premise of cloud computing is that
shared resources are cheaper than custom. But if the services are developed
in a vacuum, without conformance to any standard, the benefits are missed.
This is the challenge for retailers.
Discussion
Questions: Can retailers collaborate on common standards for critical retail
applications? Will the full potential of cloud computing be achieved without
standards?
[Author’s Commentary]
Retailers really need to step up to the plate. One of the other projects
ARTS has recently completed was a paper on SOA (Service Oriented Architecture).
This is the first step in defining a standard set of business services
that are needed by retailers. If retailers are going to achieve the benefits
of cloud computing, they need to standardize their business practices so
that different software vendors can address the requirements. This will
spur competition and give retailers service provider options.
Standards
don’t mean retailers won’t have their individual operating modes. Some
retailers will continue to use gross cost with allowances and some will
use net cost, some will be everyday low price and others will be high-low,
some will have all their sale retails end in 00 and others will use 99.
What matters is that cloud based service providers have the options in
their applications that allow retailers to achieve their goals.
I think
the potential benefits are so huge, that retailers cannot miss the opportunities
presented by cloud computing. This will make organizations like ARTS even
more important for the future of retail. A few large companies may continue
to run their own internal IT development, but the vast majority of retailers
should be looking to the cloud for IT services.
While I agree that cloud computing could benefit from some standards, and that retailers will need standards to address their specific industry requirements, I am still skeptical about whether cloud computing can be taken to any extreme–in capability, in use, in anything really. RSR, being the virtual company that it is, is effectively 100% cloud computing in its own way. Everything we do, from email to our website, is delivered over the internet. As my esteemed colleague Paula says, “It ain’t all that.” It’s not. It’s just not there yet. We deal with it because we’re small and we’re virtual, so no one wants to sign up to be tech support and have servers in their basement or garage–and truly, it doesn’t make sense for a company our size. But putting our destiny into the hands of nameless, faceless others–the reliability hasn’t been spectacular, and neither has the support.
We’re not there yet, and I think there is a long way to go, standards or not.
From a technology side, it is clear that cloud computing holds significant advantages for retailers. Not only does it expand the ability of retailers to deploy and manage their applications with very low total cost of ownership (TCO), but it also permits companies to quickly make changes that can appear across their entire store base. From a straight ROI standpoint, cloud computing would be hard to beat.
It is the culture issue that I believe will hinder retailers from deploying such a strategy in the near future. Retailers that are the most technology-forward have already deployed web architectures and are seeing the benefits of cloud-like computing. But for the rest of retailers, the “not invented here” syndrome combined with concerns about privacy will slow them down and keep them in old technology until that technology literally comes crashing down around them.
Most of the slower-to-implement companies face greater challenges internally than they do from competition or changes in customer behavior. Only when the culture adopts new values of innovation will such initiatives as “the cloud” be accepted.
Read some about cloud computing. One question–Is it safe?
The retail industry has never been credited with being an early adopter with respect to technology. I don’t expect a different reaction to cloud computing.
As is usually the case, I expect that astute, pioneering retail companies will adopt early and define the model for the majority of laggards.
It wouldn’t surprise me to see a sophisticated franchise operation like McDonald’s be one of the pioneers, given their attention to cost efficiency and the enormous scope of their IT requirements.
Cloud computing is nothing more or less than running multiple computers in parallel sequence. Owners of an iPod and a Mac for example can create a personal “cloud.” Unless the industry pulls together to establish clear–NON-HYPED–understandings of what cloud computing is and can and can’t do, we’re looking at another RFID fiasco as proponents claim it will solve all problems and opponents claim it will be the end of security. Time for the voice of reason.
Working actively in this area, I have seen it a couple of ways. I agree with another commentary that a retailer’s culture is the predominant decision maker but there are also a handful of risk taking CIOs out there who recognize the true strategic benefits of being first to market.
First off, cloud computing is a rather silly name in my opinion as it is neither descriptive nor credible to any senior decision maker who may have to bring this up to the board for approval. Outsourced services or whatever you may want to call it can and will eventually be adopted by retailers but not as a standardized offering. Each retailer defines itself in a plethora of different ways, thus each solution needs to be adaptable to the strategy of each specific retailer.
Change through evolution commonly happens; change through revolution is for the few retailers bold enough to recognize that a collaborative effort with vendor partners gives them an extended reach into the retailer community. Leveraging the intellectual capacity of the vendor’s installed base as well as the cumulative benefits and ideas of customizing each application to the strategic needs of the retailer, will result in newer ways of looking at the business.
This is where the power of the cloud excels. Minimize financial risk and greatly leverage strategic advantage.
Ryan and Mark have it nailed. Big opportunity, but big challenges. I think retail corporate cultures, and vendor hype, will both hold this back from achieving what it could.
We think cloud computing is great and we are working on a solution that will be Software as a Service Cloud Computing with Analytics tied into the Cloud. Cloud computing has many benefits as the article and others have pointed out. When you add SaaS to the equation, the benefits increase even more so. The retail industry will need to be more open about true collaboration in order for it to be successful in our industry.
There are currently several competing “cloud” applications that are different in philosophy, operation, and execution. Unless there is a group of retailers and manufacturers investing the time to determine what works for the INDUSTRY, this effort will not go far. There will be one part of the industry working with a consultant or provider creating something that works for them and proposing it as a standard. Then other groups will propose competing standards and none will be adopted.
Coming from the technology industry, I find it humorous to hear new buzz words for existing technology. “Cloud computing” is not new. It has existed for many years in different forms. The retail and CPG industry lags other industries such as banking in this area.
It’s important to distinguish between in-house versus SaaS versus hosted solutions. SaaS solutions are standardized software solutions that provide you with specific functionality. They are applications that allow very little customization and are therefore easy to manage by the vendor. Hosted solutions are software solutions that can be customized. They are offered by and managed by vendors. There is more management involved, but applications that require customization such as pricing applications, sales analytics, and business intelligence applications require customization.
Cloud computing can also be implemented “In-House,” behind your firewall. Depending on the application and the user, different solutions can be used to accomplish your goals.
There must be different “Standards” set based on whether the application is a transactional application versus an analytical application. Transactional applications are designed to “run” the business. Analytical applications are designed to “manage” the business. The models for these applications are completely different. One is designed for entry of data, the other is designed for retrieval of data. The form of “Cloud Computing” you are hoping to implement, must reflect the type of application you are using it for. The technologies that support either are very different and any standards that do not specifically detail which they are to be used for will guide companies down the wrong path and waste a lot of money.
Analytical applications need to implement technology that will provide end users with powerful tools that they need, while maintaining the ability to easily manage the application. Transactional applications do not have the same requirements. Both need to be easy to manage and many technologies exist to support either form of “cloud computing.” If you don’t first define which type of application you are creating these standards for, companies will fail in their implementations. It is up to the standards group to make sure they define this up front before making any recommendations on technology.
There is no need to fear the cloud in any way. Cloud computing is simply a new consumption and delivery model that was actually inspired by consumer internet services. Cloud computing addresses the explosive growth of internet-connected devices, and complements the increasing presence of technology in today’s world. Cloud computing is massively scalable, provides a superior user experience, and is characterized by new, internet-driven economics. It is a little more in scope that just your software residing on the internet, as opposed to your PC. Cloud computing represents a paradigm shift in the consumption and delivery of IT services.
Built on the foundation of what needs to be a “dynamic infrastructure,” I do have some reservations of how quickly retailers will not only adopt but also truly leverage this IT model. CPG companies are quickly moving to this model, enabled by their long-time adoption of ERP in their enterprises. However, with so many retailer’s infrastructures, both large and small, being comprised of a best-of-breed, point solution structure, rather than ERP, I think there will be a slower influx of cloud.
Standards are the right way to go, however, it took a long time for all retailers globally to adopt consistent UPC standards. This will most assuredly require a period of years to evolve.
Thanks Janet for clarifying the different flavors of “cloud.” You beat me to it. I agree with others that the industry standard versions of this will take some time to develop, which is very consistent with past industry standards. That said, the ability to deliver (private cloud) services into the inherently dispersed retail store organization and supply chain provide major economic opportunities. This can a combination of both the hardware and software sides. Consider these three potential levers: (1) reducing per store technology spend (multiplier), (2) reducing cost, while increasing speed and flexibility of delivering new services in store, and (3) increasing speed to integration of acquisitions and new store openings.
It is still too early to truly realize the strengths of cloud computing and the reach that it offers in a retail environment. We cannot exercise the reach that cloud computing has until we have many competitors in the cloud space who can drive several solutions to the same problem. Cloud computing is just another solution, and still not a killer application. Until we find the killer application (at speeds that everyone can embrace), cloud computing will remain a goal and not a reality.
There have been many observant and solid points made already. Cloud computing and SaaS solutions have many advantages over buying software, applicable to both ecommerce sites as well as back end systems for brick and mortar and traditional retailers.
One of the greatest advantages of this technology for ecommerce-based retailers however, is the flexibility and ease of expansion. With traditional hosting, an ecommerce retailer must buy bandwidth for their maximum expected traffic level. For much of the year, this is a waste of money as this bandwidth stands idle. With cloud computing, it is easy to contract for “as needed” bandwidth, which expands based on need. Thus, on days like today, cyber-Monday, your bandwidth can expand to enable you to handle the large volume, and contract when January comes along. This ensures that you will be able to cost-effectively address the ebb and flow of seasonal business. This alone makes it worth considering.
Cloud computing makes it possible to answer questions that you did not think you could ask. Retail analytics is an area that will benefit tremendously from cloud-computing. At Forecast Horizon, we have developed algorithms to do Price Optimization and Model Stock Optimization, which involve crunching large data-sets, and running multiple jobs in parallel. Amazon’s cloud allows us to do all this: S3 for storage, SQS to queue up and monitor the jobs, EC2 to run the optimization algorithms.
AWS is not simply an extension of a typical data-center, it is a world-class infrastructure, with components like Queuing and Storage and Computing that talk to each other, through simple APIs. Further, you only pay for the resources when you use it. In Retail Analytics, the data is mostly in passive storage. The fact that you can split up the CPU meter and the Storage meter allows you to pay for each service separately.
Having worked at a traditional enterprise software company, I have already seen the tremendous benefit that cloud-computing brings to Retail Analytics. It cuts down TCO by at least an order of magnitude. This kind of value proposition is difficult to ignore for too long.