November 6, 2017
We’re just getting started. And industry-specific solutions should be next.
Among industry pundits, there are two broad opinions about the cloud. One is that the transition is over and the cloud is here to stay: Everything that matters has moved into the cloud by now and the market is mature. The other is that cloud technology is finally becoming common in mainstream enterprise use and, although the technology is not exactly mature yet, the market is heading that way.
You could make reasonable arguments supported by evidence on either side. The “mature already” column would include reports such as RightScale’s 2017 State of the Cloud Report, which found that 85 percent of enterprises have a multi-cloud strategy and 95 percent are using the cloud in some form. The “still getting there” column would include Forrester’s prediction that public cloud expenditures will grow from $146 billion in 2017 to $236 billion in 2020.
PwC suggests a third point of view: Enterprises are just getting started with cloud technologies. As we see it, the cloud is at the end of the beginning. The groundwork has been laid, and enterprises have broadly accepted cloud services as infrastructure and for general-purpose applications.
But an enormous amount of work still must be done to build vertically tailored, cloud-based solutions on top of that stack. With that work will come enormous opportunities and massive growth, ushering in what PwC calls the “cloud-first era.”
The cloud so far
In some application categories, cloud adoption has already been extremely successful.
Web hosting: This area is well suited to the cloud, because companies tend to use a standard stack of open-source technologies that can be deployed easily, as needed, cookie-cutter style. The LAMP (Linux-Apache-MySQL-PHP) stack doesn’t have a lot of subtlety, for instance; cloud service providers can easily spin up any number of virtual servers with these and similar standard components in a predictable way.
Horizontal applications: For non-industry-specific use cases, cloud apps like those from Salesforce and Workday have been very successful. These service providers focus on use cases that apply to a wide range of companies, so they have been able to reach a broad market through a standard feature set.
Basic infrastructure: If an aspect of corporate IT infrastructure is fairly standard, there’s a good chance that cloud solutions already exist and are in relatively widespread use. The economics and broad applicability make cloud solutions an easy sell to IT decision makers: Would you spend $60 per 1TB hard drive, plus additional costs for data center management? Or, would you rather spend just $6 per month on a cloud storage service?
The next frontier: Vertical cloud solutions
Yet there is still much room for growth. InformationWeek’s 2017 State of Cloud report found that just 39 percent of IT executives expect to get 50 percent or more of their IT services from the cloud.
The growth will come from vertical, industry-specific solutions. In banking, insurance, manufacturing, and many other sectors, a multitude of specific applications haven’t been moved into the cloud yet.
For broad, generally appropriate cloud applications—like email and messaging, collaboration and conferencing solutions, and file storage—the acceptance and usage are quite high among enterprises. For example, 47 percent of companies already use cloud solutions for messaging, and just 21 percent have no plans to use the cloud. But for areas that require more vertical specialization—such as supply chain management, compliance management, and product lifecycle management—cloud penetration is much lower. For product lifecycle management, just 9 percent of companies are already using a cloud solution, while 60 percent have no plans to do so.
Cloud vendors don’t see a vertical, like manufacturing, as being a big enough business opportunity to pursue. And the up-front costs of migrating a 20- or 30-year-old legacy system into the cloud are simply too daunting for IT executives in these industries. The CIO of one manufacturing company we talked to says he has been telling cloud vendors, “Write me a check, and I’ll move to the cloud.” The costs are simply too high to do it otherwise.
You can’t just “lift and shift” something from a legacy environment into the cloud. That’s because these legacy systems were essentially handcrafted with custom data stores, specialized middleware, and more. They represent decades of embedded knowledge about how a particular industry or company works, what its partners require, how its market behaves, and so on. These legacy systems often reside on mainframes or in customized data centers that have no parallel in present-day cloud architectures.
The whole idea of a cloud data service is that the architecture is standard and can scale up or down as demand fluctuates, simply by creating or destroying new instances. Similarly, it’s easy to replicate such a service from customer to customer. Moving a highly customized system into that environment is nontrivial, as engineers say. It requires a lot of hard work, and the results are not easily transferred to other customers.
“The growth will come from vertical, industry-specific solutions. In banking, insurance, manufacturing, and many other sectors, a multitude of specific applications haven’t been moved into the cloud yet.”
It won’t be easy
We believe there’s a cloud services company with scale waiting to be born in every industry: insurance, banking, oil and gas, and more. Such industry-specific solutions will deliver the benefits of scale that only the cloud can provide, but will be adapted to the unique needs of companies in each industry.
That will require deep industry knowledge and a willingness to work closely with customers, especially in the early stages, to customize the solutions as needed.
This development hasn’t happened already because, while the potential markets are enormous, the amount of hard work and specific knowledge required are too daunting.
Look at how long it took to transform the old airline reservation system into a modern one. It took lots of heavy lifting, endless committee meetings, and many arguments. Much of this complexity had to do with business rules more than technology specifics.
The same thing will apply to creating cloud solutions for industries such as manufacturing, utilities, retail, and so on. For most companies in these sectors, the migration of their core IT infrastructure to anything new—cloud or otherwise—is a multiyear process. As a result, they tend to delay it as long as possible, patching the existing platform and making incremental improvements as long as they can.
In the insurance industry, at least three waves of prominent vendors have come and gone over the past two decades. The story is similar each time: A crop of insurance companies makes a decision to go with a vendor that’s telling a good technology story. The customers commit, but the transformations fail or take much longer than expected. Eventually, they abandon the vendor, and word of mouth kills the vendor’s prospects for new deals.
“Look at how long it took to transform the old airline reservation system into a modern one. It took lots of heavy lifting, endless committee meetings, and many arguments. Much of this complexity had to do with business rules more than technology specifics.”
More than technology
What will it take for more companies to transition into cloud applications faster and better?
This question is not just about technology. Business process transformation and culture changes may also be necessary.
Here is where we see an opportunity. In the past few years, software development has transformed from a slow, deliberative, waterfall methodology to a new, more iterative, agile approach.
The core of agile development is to break down large projects into smaller incremental upgrades. Deploying an upgrade every two weeks, collecting feedback, adjusting to changing needs in near real time, and keeping developers closely aligned with end users produces better results and larger effective changes over time than trying to overhaul everything and release a major new version every 12 or 18 months.
A similar approach could help industry-specific cloud vendors grow, while helping their customers move into the cloud. Instead of a massive, step-function change from the legacy system to a cloud system, organizations can deploy cloud technologies more incrementally, making each step far more achievable.
But to do that, IT organizations—and indeed the companies they support—will need to embrace a more agile approach.
Incremental implementation will increase success rates in delivering industry-specific cloud solutions. It will enable enormous amounts of growth in cloud technologies, as industry after industry shifts to the cloud. And it will create enormous opportunities for cloud vendors who can combine deep sector knowledge with scalable cloud technologies.
That’s why we think the cloud era is only at the end of the beginning. Yes, it’s been an enormous architectural shift. But you ain’t seen nothing yet.