Premise
Users considering public cloud for business and IT workloads should evaluate the types of use cases and workloads that have wide successful deployment in the public cloud today. Users should be more cautious about embracing workloads that are not frequently accomplished in these offerings.
In order to better understand the customers and workloads being served by public cloud providers, Wikibon examined the brief descriptions of approximately 650 customer solution stories on the Amazon Web Services (AWS) website (https://aws.amazon.com/solutions/case-studies/all/). Wikibon classified each according to 1) company type, 2) workload characteristics, 3) primary data formats, and 4) their main usage of AWS in the profile. While not conclusive, this research reveals some interesting insights about the public cloud leader’s business, and therefore about public cloud utilization generally, including:
- Cloud-based startups flock to Amazon. We all knew it, but AWS’s customer case data conclusively shows it: the digital piranhas disrupting global industries start their journeys in Amazon’s waters. In very real respects, Amazon is a key source of the emerging digital economy.
- More mature customers still tend to be cloud-based businesses. SaaS, social networks, and gaming companies lead the list of the more established brands working with Amazon. As smaller, cloud-based companies grow, they appear to be loyal to Amazon, extending their AWS relationships with additional workloads.
- The majority of traditional enterprises use AWS at the department level. The majority of AWS profiles of traditional businesses correspond to departmental-level applications (especially analytics), and not IT use cases like IaaS. Given feedback from our community, however, we suspect that AWS is under-representing IT development use cases in its profiles.
The Public Cloud is a Great Starting Point for Emerging Software and Service Providers
Of the total population of AWS customers profiled by Amazon, we classify 63% as cloud-based startups (see Figure 1). These firms are providing technology-heavy services, like mobile apps, video distribution, and social networking apps, but avoiding the upfront costs associated with building out private infrastructure. Not all of these startups are still small. AirBnB, Lyft, Pinterest, Expedia, and Spotify are all examples of cloud-based startups that started and stayed with AWS as they’ve grown. Media companies, both traditional and cloud-centric, represent 9% of the profiled AWS customers; this group primarily uses AWS for content management and distribution, including CDN. The remaining profiles (28%) feature a broad distribution of mainstream enterprises in multiple industries, including manufacturing, finance, telecom, retail, government, and healthcare.
Established AWS Companies Stick with AWS as They Scale
While known as the go-to supplier of cloud services for startups, established cloud-based companies are scaling their business on AWS across a variety of business models. Our classification shows that 38% of these companies are providing SaaS services, 24% are offering AWS-enabled business services, 10% are social network businesses, and 10% are gaming companies (see Figure 2). Smaller sub-segments include “analytics as a service” providers (9%), online retailers (4%), and payment processors (3%).
Traditional Businesses Using AWS for Departmental Apps
While the sweet spot for AWS services may be cloud-centric companies, traditional enterprises are making good use of AWS services as well. Among profiled traditional enterprises, 56% of AWS use cases are departmental or workload-specific, including marketing, engineering, and content management, i.e. they are not IT-driven.. Enterprise IT groups that buy AWS for IaaS, however, feature strongly in AWS’s enterprise customer profiles (21%), ranging from targeted offloading of storage workloads, to application development. Some even report considering AWS for a data center migration. Twelve percent of profiled enterprise AWS use cases highlighted IT groups that were using AWS for website hosting, replacing their own internally-run websites.
A stalwart 11% of traditional enterprises profiled were migrating one or more modules (i.e., billing apps, Sharepoint, order entry) from on-premises SAP, Oracle, Microsoft, Epicor and other applications to AWS. Wikibon suspects that these AWS customers are using AWS for cloud-based, inter-company processes, like payment or content sharing. The Wikibon community doesn’t report an appetite for migrating complex systems of record to AWS wholesale. However, based on AWS’s published cases, some enterprises are experimenting with how to make this work. Moreover, AWS has begun a strategic push to capture customers looking to migrate more traditional apps in the cloud, if only to counter the efforts of IBM, Oracle, and Microsoft, who seek leverage their on-premises installed bases to capture early movers. Will AWS and others make these types of complex migrations easier, and therefore less risky and costly? That is a central question to the future of the tech industry. The answer now is uncertain. But 10 years ago, most questioned AWS’s viability. As these profiles indicate, AWS tends to succeed where it dedicates resources.
Action Item
Public cloud adoption is dramatically accelerating, but many mysteries regarding adoption dynamics remain — especially regarding identifying the right workloads, the right business fit, and the right timing. Use any and all data available to start unambiguously answering these crucial what, how, and when questions for your enterprise.
Methodology
Wikibon examined the brief profiles of approximately 650 use cases provided by AWS at https://aws.amazon.com/solutions/case-studies/all/). The descriptive overviews for these customer case studies were typically 3 or 4 sentences that usually covered (or inferred) company type, primary (and often secondary) use case, data formats, and often AWS tools used and benefits derived. Wikibon classified the available data in each profile and sorted it to arrive at our findings.
We acknowledge there are obvious limitations to this sort of anecdotal analysis including:
- Description brevity – While AWS did a good job (in our view) of bottom-lining the key data points in the brief overview description of the user and their use case, skimming the lead paragraph to characterize use may not capture the entire picture
- Limits of ad hoc text analysis – Different descriptors may have been used by the profile writers to describe the characteristics. For instance “surges”, “bursts”, “massive volumes of data” may be all pointing to the same condition – or may be slightly different characteristics.
- Vendor-selected profile universe – All vendors shape their case studies to encourage prospects for their core target markets, but also to reflect their strategy and marketing goals going forward.
However, Wikibon believes this methodology and our findings provide a fair sketch of the characteristic of AWS use cases in general. Users should read the relevant cases studies from all the vendors relevant to their public cloud considerations and decide for themselves whether they pertain to their own platform decision situations.
To their credit, AWS provides perhaps a 1000 pages of customer stories on approximately 650 of its customers – at least a few of which are likely to describe common use case situations for most any user considering the public cloud for most any workload. Of course, “your mileage will vary” but AWS does provide a useful catalog of contacts a user can make to discuss the pros and cons of your target workload. Of course, one’s industry peers, analysts, and consultants should also be polled about their experiences and impressions of the offering for the particular public cloud use case you are considering.
Category Definitions
The following describes how we assigned each profiled company’s primary line of business:
- SaaS – The SaaS users appear to be looking to AWS (for the most part) to host their software without any special contribution from the IaaS side of the AWS house
- Business Services – Profiled business services users often appear to have integrated cloud technical services into their offerings as well as using AWS software hosting. This might involve challenging data formats (video, big data), managing high volume data stores, management and delivery, high speed data ingestion, and/or workloads characterized by surges and peaks – as well as high levels of scaling requirements over time.
- Social Networks – Social Networks ranged from business exchanges to purely social or affinity networks.
- Gaming – Gaming companies have massive scale, with lots of variability, so the on-demand, elastic characteristics of public cloud are a good fit.
- Analytics-as-a-Service – Analytics-as-a-Service providers supply reports, advice, consulting, etc. based on their proprietary analytic processes. Overall, analytics – adtech, recommendation engines, HPC, genomics, internet of things – appears to be a very important component of a significant minority of profiled AWS customer offerings. The fact that these use cases almost always reside outside of IT itself perhaps explains why the public cloud is such a popular landing platform for them.
- On-line Retail – This group included Retailers whose business was 100% on-line. Retailers with storefronts or other delivery channels were included in our Enterprise customer category.
- Payments Processing – E-commerce providers focusing on processing payments were included here.
- Hosting – In most cases these were web hosters, but there were a few who hosted other types of services.
We treated Media and Entertainment (9% of the profiled sample) as a separate category, lumping traditional and cloud-centric media enterprises together, because they share a strong need for content management and digital distribution.Traditional media companies will have legacy data center considerations, while internet-centric companies will not, but the core workload requirements for their public cloud expenditures make them more similar than different in our view.
