Cloud Adoption Linked to Stronger Firm Performance
By Brian Eastwood, MIT Sloan
When implementing enterprise-wide technology, executives want to be sure their investment will pay off — and rightfully so, given the time, money, and resources required to get new tech up and running. Evidence about the value of cloud-based services is often anecdotal, based on the experience of a handful of companies who have published vendor-backed case studies.
A new research paper by Wang Jin, a research fellow at the MIT Initiative on the Digital Economy, provides some of the first large-scale research on the impact of cloud adoption. The study found that transitioning to the cloud pays off, helping firms improve productivity, increase revenue, and build a foundation for initiatives such as artificial intelligence and supporting a remote workforce.
“The cloud has a positive effect on firms’ performance,” Jin said. “The effect is statistically significant and economically large, and it grows over time.”
Job listings signal the start of cloud adoption
In a 2018 paper, Jin and his co-author found that investments in the cloud were associated with higher survival rates and growth figures for small and medium-sized manufacturing firms.
Yet assessing the impact of cloud adoption on company productivity has proven difficult given a lack of large-scale empirical evidence. Part of the reason for this is practical, as most enterprises don’t capture cloud investments as a single capital expense on their balance sheet. Because cloud services can be adopted on an ad hoc basis and viewed as a flexible cost.
Part of the reason is competitive, as firms often regard migration to the cloud as a trade secret of sorts. This is especially true in the technology sector, Jin said. “You don’t want to mention who [cloud vendors] you’re working with to develop a service. That’s a high-level strategic decision that often brings a competitive edge.”
However, there’s another way to determine whether firms are adopting the cloud: Job postings. MIT economists Daron Acemoglu and David Autor demonstrated in a 2021 paper that artificial intelligence adoption impacts the skill requirements that companies listed in job postings.
Jin used this research model, measuring firms’ investment in and adoption of cloud computing through the demand for both general cloud skills (design and manufacturing) or product-specific skills (Salesforce Service Cloud or Amazon Simple Storage Service). He drew data from jobs posted through job market analytics company Burning Glass Technologies, which covered more than 200 million jobs posted since 2007.
There are two benefits to this approach, Jin said. One is that it indicates when firms truly began to “adopt” the cloud. It’s not when an official announcement is made in the SEC documentation which, even when it exists, is often a strategic move made much later than the actual adoption. Rather, it is when the first job opening seeking cloud-related skills is posted, implying a strategic commitment of cloud-related human capital investment. Relatedly, it also affirms that a firm’s cloud strategy is equal parts an investment in technology and complementary skilled labor.
“One of the biggest barriers to cloud adoption, especially in large firms, is a lack of cloud-related skills,” Jin said. This is backed by an Accenture report.
Cloud benefits: revenue growth, lower operating expenses, AI adoption
Jin found that revenue growth for public firms adopting cloud technology was 2.3% to 6.9% higher than for non-adopters in the last decade, the midpoint being when the firm first posted its first job seeking cloud-related skills.
In addition, by estimating the business value of cloud computing, Jin concluded that the hiring of each worker with cloud-related skills generates around $1.4 million in sales.
If you compare two firms, and one is cloud-ready and one isn’t, the first is going to continue to perform better.
Several factors could contribute to this growth. The cloud’s flexible deployment model gives enterprises the option to start small, typically within a single business unit, and to increase or decrease scale as desired. This pay-as-you-go model is a stark contrast to traditional on-premises software applications that require years of planning and still fail at alarming rates — up to 75% for enterprise resource planning, for example.
“With the cloud, you can start as an experiment, make sure that you can make it work, and then roll it out across the larger organization,” Jin said. “And if something fails, you wouldn’t lose the fixed costs of IT infrastructure.”
The shift away from on-premises IT infrastructure also eliminates geographic restrictions on where firms base both technology assets and the employees using them. This lowers operating expenses, as offices for knowledge workers don’t need to be in high-cost markets, and it expands the pool of job candidates with cloud skills.
Finally, cloud adoption for many firms serves as the starting point for artificial intelligence and machine learning initiatives. Jin said it’s important to note that cloud adoption and AI and machine learning adoption are interdependent. The cloud is often a foundation for AI, he added.
“If you want to run large-scale models, it’s natural to have to run them on the cloud given the demand for computing power,” Jin said.
Cloud challenges: organization and workflow changes
Companies also face challenges associated with cloud adoption.
For starters, initial investments can be difficult to predict, Jin said. As suggested by the productivity J-curve, there are “hidden intangible” costs associated with any technology deployment, and the cloud is no exception.
“You need the labor and the skills to make it work better,” Jin said. “These complementary investments take time. Certain firms might even have some short-term losses.”
The transition to the cloud also upsets the traditional way of working. Even amid the recent shift to remote work, most organizational structures and technology use cases center on in-person work using on-premises IT systems. If these rigid business practices remain in place as a firm adopts cloud-based services, the resulting misalignment may nullify workers’ productivity gains and stifle growth, Jin said.
The utility industry provides a telling example. In traditional workflows, employees go into the field to read meters and log numbers in a notebook. When firms adopt cloud-based services, this work is done using a mobile app.
“These workers have been doing the same exercise for 30 years. There’s a process of upskilling that needs to happen,” Jin said. “You can buy the technology, but you need people to operate it. They need to have cloud-related skills.”
Despite these challenges, Jin said cloud adoption has proven its worth, especially during the pandemic.
In a separate paper, Jin and his coauthors found that firms that had the technology in place to enable remote work prior to the pandemic fared better than their competitors in terms of higher sales, net incomes, and stock market returns. “If you need to allow employees to work remotely, you need the right infrastructure to do so — and the cloud is one of the fundamental infrastructure needs,” Jin said.
Not surprisingly, cloud adoption has also accelerated through the pandemic. A survey from Palo Alto Networks pointed to a 25% increase in cloud usage in 2021.
For many of these new adopters, Jin said the initial benefits may be difficult to measure. These firms haven’t had the time to make complementary investments, while their cloud strategies have largely emphasized maintaining operations.
Over time, though, the effect will be felt. “Firms that were better prepared performed better during the pandemic. That has significant implications for market structures,” Jin said. “If you compare two firms, and one is cloud-ready and one isn’t, the first is going to continue to perform better. That is likely to have a long-lasting effect — and it could restructure the market.”
Originally published at https://mitsloan.mit.edu on June 14, 2022.