Following continuous software engineering practices, there has been an\nincreasing interest in rapid deployment of machine learning (ML) features,\ncalled MLOps. In this paper, we study the importance of MLOps in the context of\ndata scientists' daily activities, based on a survey where we collected\nresponses from 331 professionals from 63 different countries in ML domain,\nindicating on what they were working on in the last three months. Based on the\nresults, up to 40% respondents say that they work with both models and\ninfrastructure; the majority of the work revolves around relational and time\nseries data; and the largest categories of problems to be solved are predictive\nanalysis, time series data, and computer vision. The biggest perceived problems\nrevolve around data, although there is some awareness of problems related to\ndeploying models to production and related procedures. To hypothesise, we\nbelieve that organisations represented in the survey can be divided to three\ncategories -- (i) figuring out how to best use data; (ii) focusing on building\nthe first models and getting them to production; and (iii) managing several\nmodels, their versions and training datasets, as well as retraining and\nfrequent deployment of retrained models. In the results, the majority of\nrespondents are in category (i) or (ii), focusing on data and models; however\nthe benefits of MLOps only emerge in category (iii) when there is a need for\nfrequent retraining and redeployment. Hence, setting up an MLOps pipeline is a\nnatural step to take, when an organization takes the step from ML as a\nproof-of-concept to ML as a part of nominal activities.\n