ClearPath Research Methodology

This survey was conducted and produced by ClearPath Strategies (www.clearpath-strategies.com), a strategic consulting and research firm for the world’s leaders and progressive forces. Following are the firm’s research notes for this survey.

Respondent selection

Each report includes information on number of respondents. Respondents are sourced from a leading global online panel provider. They were selected from the panel based on geographic and role-based quotas, as well as screening questions based on role in IT, decision-making role, company size, and how long they have been in IT. Selected respondents were further screened based on self-reported IT knowledge and attentiveness to survey questions.

Role quotas

The survey divided respondents into four broad IT “roles”: Developer 30%, Operations 30%, Manager 20%, and Line of business leadership 20%. Respondents were asked to select which role—from a list of nine options—most closely described their primary responsibility, even if no one was quite right or even if they performed more than one of these roles. Answers were consolidated into those four broad roles.

Geographic quotas

The survey included respondents from the US (18%), UK (17%), China (10%), Japan (7%), Germany (17%), Canada (15%), India (11%), and South Korea (6%). We combine these broadly into three regions: North America (33%; US and Canada), Europe (33%; UK and Germany), and Asia (33%; China, Japan, India, and South Korea).

Industry

Although no industry-level quotas were deployed, we monitored the data to ensure that no single industry was over-represented in the data. The final breakdown of respondents by industry is as follows: IT (software, hardware, services) 28%, Manufacturing 14%, Financial services 12%, Government 6%, Health care 5%, Business services 5%, Education 4%, Telecommunications/ISP/Web hosting 4%, Transportation and logistics 4%, Retail 3%, Construction/engineering 3%, Utilities 2%, Wholesale 2%, Consumer services 1%, Life sciences 1%, Mining and natural resources 1%, Non-profit 1%, and Other 2%.

Respondent screens

Potential respondents were screened out on several criteria:

  • Role: All respondents who selected either “IT professional / Support / Help Desk—I provide general HW and SW support to non-IT staff” or “Non-IT professional—I work in a non-IT division and am not responsible for IT decisions in my line of business” were excluded from the survey.
  • Company size: All respondents must self-report that their companies have minimum 100 employees. All potential respondents from smaller companies were excluded. In total, the survey includes 40% from companies with 100-999 employees, 32% from companies with 1,000 to 9,999 employees, 19% from companies with 10,000 to 99,999 employees, and 10% from companies of 100,000 or more employees.
  • Time in IT: Respondents must have spent minimum 2 years working in or with IT in order to qualify for the survey. In total, 63% of respondents have spent more than 10 years in IT, with 37% having spent 2-10 years.
  • Information level: In our experience, it is possible to have “qualifying respondents” who nevertheless prove to have too little information or knowledge about the space to provide useful data from which to draw insights. We therefore apply an “information” screen to respondents as well. Specifically, we ask whether or not respondents could explain certain terms to their colleagues, if asked to do so. In order to qualify for this survey, a respondent must say “yes” to this question for both the terms “cloud computing” and “virtualization.” There were a few exceptional respondents allowed in who did not say yes to both of these questions, but over 99% of respondents in the data said yes to both.
  • “Attention” level: It is easy for respondents to speed through surveys or not pay enough attention to provide useful data. We make an effort to exclude these respondents as well, as they provide generally less useful data. In this survey, respondents were screened out for “attention” reasons if they said they could explain the made-up term “Greenfield as a Service (GaaS)” to a colleague in the same question used for the Information Screen noted above. Additionally, respondents were excluded if—when answering the “PaaS awareness question”—they selected the logo/company names for McDonald’s or Starbucks. Both of those companies were included in the list of PaaS products, and respondents were asked only to select the PaaS offerings they were familiar with and specifically instructed not to select non-PaaS offerings.

Definitions

An important finding from our GPS studies to date—both qualitative and quantitative—is that IT decision makers have a much less clear-cut vision of the technology stack and do not always make clean distinctions between IaaS and PaaS as we might. For example, although we might not ourselves describe AWS, Azure, or Kubernetes as being true “PaaS” or might say they have “PaaS-like features,” the respondents may have a different perspective. Therefore, in the survey, the following definitions were used for PaaS and Containers:

  • PaaS: “PaaS—also called “Cloud Application Platform” or simply “Cloud Platform”—is a category of cloud computing services that provides a platform allowing companies to develop, run, and manage enterprise and web applications without the complexity of building and maintaining the infrastructure and middleware typically associated with developing and launching an app.”
  • Containers: “Containers offer a way to virtualize an operating system, isolating processes and providing limited visibility and resource utilization—such that the process appears to be running on a separate machine. Containers can allow applications to be packaged, with all of its dependencies, into a single, standardized unit.”

A note on margin of error

It is technically impossible and improper to list a margin of error for a survey of this type. The respondents for this sample were drawn from an online panel with an unknown relationship to the total universe, about which we also do not know the true demographics. As such, the exact representativeness of this, or any similarly produced sample, is unknown