Data science and machine learning are trending upward even while many areas of the economy face headwinds due to COVID-19.
Per a recent survey conducted by Wing, a leading venture capital firm that focuses on enterprise technology, companies plan to invest more in data science and machine learning in 2021 — and, in some cases, significantly more.
From Oct. 20 – Oct. 28, 2020, Wing surveyed 88 of the “senior-most data scientists at global corporations and venture-backed startups.” Each of the respondents leads his/her company’s data science department. Here are some results:
Even against the backgrop of COVID, 48% of respondents at public companies indicated that DS budgets will grow at 25% or greater next year. Even more impressive is that 45% of respondents at private companies indicated DS budget growth of 50% or greater next year pic.twitter.com/oJuhCbBR82
— Jake Flomenberg (@jflomenb) November 2, 2020
As Wing partner Jake Flomenberg notes, nearly half of all public companies which responded to the survey indicated that their data science budgets would grow by at least 25 percent in 2021. Only 26 percent of those companies reported no planned increase, meaning that 74 percent plan to increase their data science/machine learning budgets by at least 10 percent.
Among private companies that responded, 45 percent plan to expand their budgets by at least 50 percent, while 15 percent are planning to double their budgets for 2021.
In addition, Wing’s survey revealed the specific technological improvements these companies see emerging in the near future. In the next 1-2 years, they ranked the most important developments for data science and machine learning as such:
When the question is framed in terms of needed technology development for the ML lifecycle rather than challenges, AutoML edges its way into the top 3 alongside explainability and MLOps pic.twitter.com/pIJs4iDS3D
— Jake Flomenberg (@jflomenb) November 2, 2020
“The majority of the effort is going to shift to monitoring and maintaining, rather than just the earlier stages of data labeling and processing,” one data science leader told Wing when discussing the need for development in MLOps (Machine Learning Operations).
An enterprise software leader added to that sentiment, saying, “You need to monitor and update daily. You would think it’s 10% of the effort, but it takes 40% in reality.”
View the full results of Wing’s survey here.
