{"id":733981,"date":"2021-03-24T10:03:59","date_gmt":"2021-03-24T17:03:59","guid":{"rendered":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/?p=733981"},"modified":"2022-11-01T09:00:25","modified_gmt":"2022-11-01T16:00:25","slug":"advancing-organizational-science-using-network-machine-learning-to-measure-innovation-in-the-workplace","status":"publish","type":"post","link":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/blog\/advancing-organizational-science-using-network-machine-learning-to-measure-innovation-in-the-workplace\/","title":{"rendered":"Advancing organizational science using network machine learning to measure innovation in the workplace"},"content":{"rendered":"\n<figure class=\"wp-block-image alignwide size-large\"><img decoding=\"async\" src=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/1400x788_WTI_still_no_logo-scaled.jpg\" alt=\"Machine Learning Graph\"\/><\/figure>\n\n\n<aside id=accordion-557fd47a-19e6-4175-b3c1-7db3d39c515c class=\"msr-table-of-contents-block accordion mb-5 pb-0\" data-bi-aN=\"table-of-contents\">\n\t<button class=\"btn btn-collapse bg-gray-100 mb-0 display-flex justify-content-between\" type=\"button\" data-mount=\"collapse\" data-target=\"#accordion-collapse-557fd47a-19e6-4175-b3c1-7db3d39c515c\" aria-expanded=\"true\" aria-controls=\"accordion-collapse-557fd47a-19e6-4175-b3c1-7db3d39c515c\">\n\t\t<span class=\"msr-table-of-contents-block__label subtitle\">In this article<\/span>\n\t\t<span class=\"msr-table-of-contents-block__current mr-4 text-gray-600 font-weight-normal\" aria-hidden=\"true\"><\/span>\n\t<\/button>\n\t<div id=\"accordion-collapse-557fd47a-19e6-4175-b3c1-7db3d39c515c\" class=\"msr-table-of-contents-block__collapse-wrapper collapse show\" data-parent=\"#accordion-557fd47a-19e6-4175-b3c1-7db3d39c515c\">\n\t\t<div class=\"accordion-body bg-gray-100 border-top pt-4\">\n\t\t\t<ol class=\"msr-table-of-contents-block__list\">\n\t\t\t\t\t\t\t\t\t<li class=\"msr-table-of-contents-block__list-item\">\n\t\t\t\t\t\t<a href=\"#using-network-machine-learning-to-measure-collaboration-changes-through-covid-19\" class=\"msr-table-of-contents-block__list-item-link\">Using network machine learning to measure collaboration changes through COVID-19<\/a>\n\t\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t<li class=\"msr-table-of-contents-block__list-item\">\n\t\t\t\t\t\t<a href=\"#a-comparison-of-workplace-siloing-in-the-united-states-and-new-zealand\" class=\"msr-table-of-contents-block__list-item-link\">A comparison of workplace siloing in the United States and New Zealand<\/a>\n\t\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t<li class=\"msr-table-of-contents-block__list-item\">\n\t\t\t\t\t\t<a href=\"#instilling-an-innovation-mentality-and-deepening-inter-team-collaboration-going-forward\" class=\"msr-table-of-contents-block__list-item-link\">Instilling an innovation mentality and deepening inter-team collaboration going forward<\/a>\n\t\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t<li class=\"msr-table-of-contents-block__list-item\">\n\t\t\t\t\t\t<a href=\"#a-note-on-privacy\" class=\"msr-table-of-contents-block__list-item-link\">A note on privacy<\/a>\n\t\t\t\t\t<\/li>\n\t\t\t\t\t\t\t<\/ul>\n\t\t<\/div>\n\t<\/div>\n\t<span class=\"msr-table-of-contents-block__progress-bar\"><\/span>\n<\/aside>\n\n\n\n<p>Is innovation another loss due to the global COVID-19 pandemic? Indicators reveal challenges to overcome\u2014as well as opportunities to build on our <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/insights.office.com\/workplace-analytics\/uncovering-resilience-measuring-organizational-networks-during-crisis\/\">collective experience gained<span class=\"sr-only\"> (opens in new tab)<\/span><\/a> in the last year. <a href=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/publication\/workgroup-mapping-visual-analysis-of-collaboration-culture\/\">Measuring collaboration using network machine learning<\/a> provides a powerful diagnostic tool to address issues as they arise. One challenge lies in how we build relationships in the workplace. With many transitioning to work-from-home environments, more flexible schedules, and new methods of working, there is evidence that interactions at work have been impacted in multifaceted ways.<\/p>\n\n\n\n<div class=\"annotations \" data-bi-aN=\"margin-callout\">\n\t<article class=\"annotations__list card depth-16 bg-body p-4 annotations__list--right\">\n\t\t<div class=\"annotations__list-item\">\n\t\t\t\t\t\t<span class=\"annotations__type d-block text-uppercase font-weight-semibold text-neutral-300 small\">Report<\/span>\n\t\t\t<a href=\"https:\/\/cm-edgetun.pages.dev\/en-us\/worklab\/work-trend-index\/hybrid-work\" data-bi-cN=\"Work Trend Index \" data-external-link=\"false\" data-bi-aN=\"margin-callout\" data-bi-type=\"annotated-link\" class=\"annotations__link font-weight-semibold text-decoration-none\"><span>Work Trend Index <\/span>&nbsp;<span class=\"glyph-in-link glyph-append glyph-append-chevron-right\" aria-hidden=\"true\"><\/span><\/a>\t\t\t\t\t<\/div>\n\t<\/article>\n<\/div>\n\n\n\n<p>This year&#8217;s <a href=\"https:\/\/cm-edgetun.pages.dev\/en-us\/worklab\/work-trend-index\/hybrid-work\">Work Trend Index<\/a> shows that the world&#8217;s information workers have begun a fundamental transition into a new world of work. According to the report, over 70 percent of global workers surveyed want flexible, remote work options to continue. At the same time, over 65 percent say they want more in-person time with teams. This future is dependent on utilizing new technologies for communication and collaboration, meeting the needs of individuals when it comes to motivation and mental health, and evolving ideas to accommodate modern workplaces\u2014big and small.<\/p>\n\n\n\n<p>As the world over the past year has changed in the wake of COVID-19, employees\u2019 networks have narrowed. It\u2019s important that we find ways to rebuild our wider work networks as we adjust to new ways of collaborating with others. Innovative teams have been shown to frequently have access to novel information from across their organization and to create new collaborations on a regular basis. Workers will need to tap into a diversity of ideas and viewpoints to foster creativity. The bottom line is that social ties and support networks are important to both our productivity and our ability to adapt at work. You can learn more about some of the takeaways for business leaders and workplaces in this <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/hbr.org\/2021\/03\/what-a-year-of-wfh-has-done-to-our-relationships-at-work\">Harvard Business Review article<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>. <\/p>\n\n\n\n<div class=\"annotations \" data-bi-aN=\"margin-callout\">\n\t<article class=\"annotations__list card depth-16 bg-body p-4 annotations__list--right\">\n\t\t<div class=\"annotations__list-item\">\n\t\t\t\t\t\t<span class=\"annotations__type d-block text-uppercase font-weight-semibold text-neutral-300 small\">Code<\/span>\n\t\t\t<a href=\"https:\/\/github.com\/microsoft\/graspologic\" data-bi-cN=\"Graspologic: python package for graph statistics\" data-external-link=\"false\" data-bi-aN=\"margin-callout\" data-bi-type=\"annotated-link\" class=\"annotations__link font-weight-semibold text-decoration-none\"><span>Graspologic: python package for graph statistics<\/span>&nbsp;<span class=\"glyph-in-link glyph-append glyph-append-chevron-right\" aria-hidden=\"true\"><\/span><\/a>\t\t\t\t\t<\/div>\n\t<\/article>\n<\/div>\n\n\n\n<p>Within Microsoft Research, we are developing network machine learning to derive insights from connected data. In the case of the Work Trend Index, this technique can be applied to understand global patterns across Microsoft 365. Our Graph AI efforts have led us to develop tools and techniques to use connected data to service a wide variety of research application areas, such as <a href=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/publication\/using-web-scale-graph-analytics-to-counter-technical-support-scams\/\">tech fraud in the Microsoft Digital Crimes Unit<\/a>, <a href=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/publication\/making-sense-of-search-using-graph-embedding-and-visualization-to-transform-query-understanding\/\">search in Microsoft Bing<\/a>, and others. The <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/github.com\/microsoft\/graspologic\">graspologic open-source Python package<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>, developed jointly with Johns Hopkins University performs partitioning, visual layouts, and network embeddings based on our research methods. Recently, we&#8217;ve been making advances in applying network machine learning to inform new solutions across Microsoft 365 and <a href=\"https:\/\/cm-edgetun.pages.dev\/en-us\/microsoft-viva\/\">Microsoft Viva<\/a>. In order to advance both theoretical research in Graph AI and applied research in organizational science, we have formed collaborations with research organizations, including Johns Hopkins University Department of Applied Mathematics and Statistics and the Michael G. Foster School of Business at the University of Washington.<\/p>\n\n\n\n<p>By using AI to understand the connection between employees, teams, and organizations, our research enables a deeper understanding of how the workplace has changed over the last year, in particular how the interactions across and within groups of information workers have been impacted. Using network partitioning, visual mappings of organizational networks, and graph metrics, we seek to gain insights into workplace communication trends within Microsoft 365.<\/p>\n\n\n\n<h2 id=\"using-network-machine-learning-to-measure-collaboration-changes-through-covid-19\">Using network machine learning to measure collaboration changes through COVID-19<\/h2>\n\n\n\n<p>Executives want to understand how they can better support the evolving communication and collaboration needs of their employees using data-driven, privacy-protected analyses. Microsoft 365 gives us a lens into the <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/insights.office.com\/networks\/toward-resilience-adapting-to-crisis-through-the-lens-of-organizational-networks\/\">patterns of employee collaboration<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>. For this study, we analyzed a subset of email interactions and meeting interactions in Microsoft Teams and Outlook. This aggregated data removes personal or organization-identifying information, such as company name. At Microsoft, we take <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/privacy.microsoft.com\/en-us\">privacy<span class=\"sr-only\"> (opens in new tab)<\/span><\/a> seriously. <a href=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/blog\/advancing-organizational-science-using-network-machine-learning-to-measure-innovation-in-the-workplace\/#toc-heading-3\">Learn more about our approach.<\/a> <\/p>\n\n\n\n<p>As much of the world shifted to work from home in response to COVID-19, we immediately saw a massive increase in Teams meetings and a modest increase in email volume (see Figure 1 below). Both these increases persist as we cross the one-year mark. As we explore more of the trends in these findings, it\u2019s worth keeping in mind that this was not a typical year for information workers and workplaces. Many people did not elect to work remotely, were working in less-than-ideal work spaces, and were attempting to balance work with personal obligations, like childcare.<\/p>\n\n\n\n<figure class=\"wp-block-image aligncenter size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"562\" src=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/EmailMeetingVolume_plot_Xs_on_axes103173-1024x562.png\" alt=\"A line graph shows email and meeting increases from January 2019 to January 2021. Meetings increased at a very slow rate from January 2019 to about February of 2020, where they increased substantially  throughout an adjustment period after COVID-19 became widespread. There was a slight flattening from July to August 2020, and then meetings shot upward again. They taper and decrease in the winter of 2020, but increase again approaching January 2021.\n\nEmails show an up and down trend over time, with a pattern of steep rises that are far greater than the drops. Emails rise from February 2019 to May, then dip before rising again in June. There are a series of sharp ups and downs between July 2019 and March of 2020, ultimately resulting in a steep increase in email to a peak in March. During the adjustment period, meeting volume decreased through the spring and then rose, ultimately resulting in a decrease in email by July 2020, still much higher than the previous year. There is a sharp rise and then decrease in fall, with a final increase in December of 2020, ending just below the highest rate reached in March 2020. \" class=\"wp-image-735487\" srcset=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/EmailMeetingVolume_plot_Xs_on_axes103173-1024x562.png 1024w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/EmailMeetingVolume_plot_Xs_on_axes103173-300x165.png 300w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/EmailMeetingVolume_plot_Xs_on_axes103173-768x422.png 768w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/EmailMeetingVolume_plot_Xs_on_axes103173-1536x843.png 1536w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/EmailMeetingVolume_plot_Xs_on_axes103173-2048x1124.png 2048w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/EmailMeetingVolume_plot_Xs_on_axes103173-16x9.png 16w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><figcaption>Figure 1: Total email and meeting volume over time included in this study. The five-month period from March 2020 through July 2020 is considered an adjustment period as information workers shifted to working from home.<\/figcaption><\/figure>\n\n\n\n<p>In a communication graph, we identify workgroups based on collaboration patterns to understand the relationships between elements of an organization. While these workgroups frequently overlap with the official organizational chart, they instead represent the informal connections through which work actually gets done.<\/p>\n\n\n\n<p>We use <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/graspologic.readthedocs.io\/en\/latest\/reference\/partition.html#leiden-and-hierarchical-leiden\">hierarchical Leiden clustering<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>, implemented in graspologic, to automatically identify workgroups based on their communication patterns. <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/www.nature.com\/articles\/s41598-019-41695-z\/\">Leiden clustering<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>, a recent method for finding clusters in networks, was chosen due to its speed, when working with large amounts of data, and its performance improvements over previous methods. We\u2019ve used <a href=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/publication\/trimming-the-hairball-edge-cutting-strategies-for-making-dense-graphs-usable\/\">learnings from prior visualization research<\/a> to construct a network map that represents the organization.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"1024\" src=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Microsoft_rotated103248-1024x1024.png\" alt=\"A data visualization showing Microsoft employees represented as nodes. This is an abstraction of different groups of employees, where the more dense circles show those who communicate more often. One takeaway is that there are definitive small, dense circles that show colleagues within specific workgroups interacting more often.\" class=\"wp-image-735523\" srcset=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Microsoft_rotated103248-1024x1024.png 1024w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Microsoft_rotated103248-300x300.png 300w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Microsoft_rotated103248-150x150.png 150w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Microsoft_rotated103248-768x768.png 768w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Microsoft_rotated103248-12x12.png 12w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Microsoft_rotated103248-180x180.png 180w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Microsoft_rotated103248-360x360.png 360w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Microsoft_rotated103248.png 1429w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><figcaption>Figure 2: This image depicts Microsoft, with each node representing an employee. These nodes appear in proximity to other nodes they communicate with most closely. We include a subset of de-identified US-based employees. Workgroups have been derived using network machine learning and are denoted by color.<\/figcaption><\/figure>\n\n\n\n<p>The layout was automatically generated with graspologic. We use a generalized node2vec embedding, UMAP for dimensionality reduction to 2D space, and finally an occlusion removal algorithm. The colors were calculated using network machine learning and represent workgroups that communicate more with each other than with the rest of the company. Once we have identified the workgroups and generated a layout, we calculate analytics at the workgroup and organization levels to preserve employee privacy while providing executives with actionability. <\/p>\n\n\n\n<div class=\"wp-block-media-text has-vertical-margin-small  has-vertical-padding-none  is-stacked-on-mobile has-light-gray-background-color has-background\"><figure class=\"wp-block-media-text__media\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"241\" src=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Figure-2B_WTI-blog-1024x241.png\" alt=\"Diagram\" class=\"wp-image-735064 size-full\" srcset=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Figure-2B_WTI-blog-1024x241.png 1024w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Figure-2B_WTI-blog-300x71.png 300w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Figure-2B_WTI-blog-768x181.png 768w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Figure-2B_WTI-blog-16x4.png 16w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Figure-2B_WTI-blog.png 1430w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure><div class=\"wp-block-media-text__content\">\n<p class=\"has-normal-font-size\"><strong>Words to know: Workgroup siloing<\/strong>. Workgroup siloing, also known as modularity, helps us identify if a network is more connected or more siloed and provides a score to rate how siloed the network is. The higher this score, the more segmented the group of nodes are\u2014as seen in the figure.<\/p>\n<\/div><\/div>\n\n\n\n<p>From April 2019 to April 2020, modularity, a measure of workgroup siloing, rose around the world. As more people moved to remote work, there was a tendency to communicate more among familiar people and less across workgroups. The year over year increase in modularity was found to be statistically significant. This small overall change still indicates a sizable impact on individual workgroups and organizations.<\/p>\n\n\n\n<p>Using modularity as a measure of workgroup siloing, we can look at a set of organizations to observe a large increase in modularity around the time work-from-home orders were being issued worldwide. We are currently researching the increasing modularity to help leaders understand how to respond in light of these significant changes.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"948\" height=\"625\" src=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Modularity_plot102513.png\" alt=\"Diagram\" class=\"wp-image-735502\" srcset=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Modularity_plot102513.png 948w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Modularity_plot102513-300x198.png 300w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Modularity_plot102513-768x506.png 768w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Modularity_plot102513-16x12.png 16w\" sizes=\"auto, (max-width: 948px) 100vw, 948px\" \/><figcaption>Figure 3: Mean modularity versus time for 2019 and 2020. The shaded areas around the red and blue lines indicate one standard error. Note that modularity began to increase in March of 2020 and remains elevated above 2019 levels.<\/figcaption><\/figure>\n\n\n\n<h2 id=\"a-comparison-of-workplace-siloing-in-the-united-states-and-new-zealand\">A comparison of workplace siloing in the United States and New Zealand<\/h2>\n\n\n\n<p>This increased workgroup siloing has persisted into 2021. We consider the United States and New Zealand to understand the effect of society reopening on modularity. United States information workers have not yet returned to the office on a large scale, while New Zealand has eased restrictions several times. Modularity remains elevated in the United States; in New Zealand, after October\u2019s move to Alert Level 1 (see Figure 4), we see modularity gradually decrease. When lockdowns are relaxed, we see that teams become more connected. This trend is also observed in other countries. Of some concern, however, New Zealand has yet to return to pre-pandemic levels of modularity, indicating a persistent effect to date.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"905\" src=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/US_NZ_plot102512-1024x905.png\" alt=\"Diagram\" class=\"wp-image-735505\" srcset=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/US_NZ_plot102512-1024x905.png 1024w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/US_NZ_plot102512-300x265.png 300w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/US_NZ_plot102512-768x679.png 768w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/US_NZ_plot102512-1536x1357.png 1536w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/US_NZ_plot102512-14x12.png 14w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/US_NZ_plot102512.png 1682w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><figcaption>Figure 4: Workgroup siloing in 2020 in the United States and New Zealand. Both the United States and New Zealand still have elevated levels of workplace siloing into 2021. In the figure, New Zealand has specific dates coinciding with <a href=\"https:\/\/covid19.govt.nz\/alert-system\/\">centralized policy<\/a> as a result of COVID-19. The United States took a localized approach, so dates are not reflected.<\/figcaption><\/figure>\n\n\n\n<div class=\"wp-block-media-text has-vertical-margin-small  has-vertical-padding-none  is-stacked-on-mobile has-light-gray-background-color has-background\"><figure class=\"wp-block-media-text__media\"><img loading=\"lazy\" decoding=\"async\" width=\"907\" height=\"618\" src=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Figure-3B_WTI-blog.png\" alt=\"chart, radar chart\" class=\"wp-image-735070 size-full\" srcset=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Figure-3B_WTI-blog.png 907w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Figure-3B_WTI-blog-300x204.png 300w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Figure-3B_WTI-blog-768x523.png 768w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Figure-3B_WTI-blog-16x12.png 16w\" sizes=\"auto, (max-width: 907px) 100vw, 907px\" \/><\/figure><div class=\"wp-block-media-text__content\">\n<p class=\"has-normal-font-size\"><strong>Words to know: Workgroup stability. <\/strong>Workgroup stability measures the change in workgroups from month to month. Specifically, it calculates the similarity between two sets of data clusters using the adjusted rand index. The clusters are the workgroups of employees, calculated using machine learning and denoted by color in the image. A higher workgroup stability score means that communication patterns are more stable and more like the prior month.<\/p>\n<\/div><\/div>\n\n\n\n<p><em>Workgroup stability,<\/em> which measures the change in workgroup membership over time, varies by the type of communication being used. Email, a more formal method of communication, declines in workgroup stability immediately after COVID-19 becomes widespread. Stability remains below 2019 throughout the adjustment period. After this period of adjustment, we see email stability return to 2019 levels. The more informal communication medium of Teams meetings saw a sudden rise in stabilization over the adjustment period continuing into 2021.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"594\" src=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Stability_plot102473-1024x594.png\" alt=\"Workgroup stability for meetings and emails rom January 2019 to January 2021. The scale for workgroup stability goes from 0 to 0.3, and then up in increments of tenths to 0.7 and above. \n\nMeetings workgroup stability starts around 0.2 in February 2019, dipping slightly before rising gradually to above .3 in February 2020. They drop until March at the beginning of an adjustment period, where stability increases sharply above 0.5, then levels off at May 2020. Stability stays relatively level until September 2020, where it then ticks upward before leveling off between 0.5 and 0.6 for the last quarter of the year. \n\nEmails workgroup stability begins in March 2019 above 0.7, gradually increasing and then decreasing to 0.7 at the end of July 2019. Stability dips slightly before rising from August to November, where it settles in at a high above 0.7. When the adjustment period starts in March 2020, stability decreases gradually, with a slight increase between April and June, ultimately going down to about 0.7. at the end of August 2020. From August to October, there is an increase before a leveling off. The end result is a stability that is higher than at the beginning of March 2019. \n\" class=\"wp-image-735511\" srcset=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Stability_plot102473-1024x594.png 1024w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Stability_plot102473-300x174.png 300w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Stability_plot102473-768x446.png 768w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Stability_plot102473-1536x892.png 1536w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Stability_plot102473-2048x1189.png 2048w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/Stability_plot102473-16x9.png 16w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><figcaption>Figure 5: Workgroup stability over time for emails and meetings. Meeting stability continued to rise even after the increase in volume leveled out.<\/figcaption><\/figure>\n\n\n\n<p>We see a large increase in Microsoft Teams usage beginning in March, as organizations adapted to many employees working from home. This increase in meeting workgroup stability likely is the effect of regular workplace interactions moving to Teams. Rather than an occasional way to connect, Teams has become a necessary tool for daily interaction with work colleagues.<\/p>\n\n\n\n<p>Circling back, the findings in the Work Trend Index reveal there are some positive takeaways from the solidifying of existing connections. For example, 39 percent of people surveyed said that they feel like they \u201care more likely to be their full, authentic selves at work.\u201d There were co-workers meeting one another\u2019s families or pets for the first time virtually and other signs that people were leaning on their close colleagues for support. You can learn more about those findings in the <a href=\"https:\/\/cm-edgetun.pages.dev\/en-us\/worklab\/work-trend-index\/hybrid-work\">Work Trend Index.<\/a><\/p>\n\n\n\n<h2 id=\"instilling-an-innovation-mentality-and-deepening-inter-team-collaboration-going-forward\">Instilling an innovation mentality and deepening inter-team collaboration going forward<\/h2>\n\n\n\n<p>As we look ahead to a return to the workplace, whether that means a hybrid or full return, organizations and their employees will be able to take advantage of in-person opportunities to build community. In an interview with Harvard Business Review, Dr. Nancy Baym, Senior Principal Researcher at Microsoft, recommends making space for informal employee connections. \u201cThese types of interactions are critical to a thriving workplace\u2014for employees and organizations. They spur spontaneous and one-off conversations, especially with people outside our immediate team, expand our knowledge and can energize our thinking.\u201d<\/p>\n\n\n\n<p>Microsoft Researchers and collaborators are conducting further studies and are committed to using findings like these to create new solutions as we navigate the changes unfolding in the workplace, resulting in a shift in how businesses and individuals perceive and accomplish work goals. Check out the <a href=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/project\/the-new-future-of-work\/\">The New Future of Work project page<\/a> to explore some of the ways researchers have been developing new methods for collaboration, productivity, and optimization of tools for communication.<\/p>\n\n\n\n<p>Returning to our original question of the state of innovation, new ideas are often correlated with access to novel information. We observed that the workgroup siloing score for both Teams and emails rose around the time work-from-home orders were issued. Workgroups became more siloed, which could lead to higher efficiency and productivity. However, this is also likely at the expense of discovering new ideas that can lead to innovative thinking. This could also mean that inter-team communications aren\u2019t as strong now as they were before. Since March, meetings became more stable, while email did not. Meeting stability increased as daily workplace interaction shifted online. We continue to investigate these metrics.<\/p>\n\n\n\n<p>The shifts in the last year to predominantly looking inward will make it more difficult to discover the new ideas that lead to innovation. In the past, these connections were facilitated by watercooler chats and hallway conversations. As we slowly emerge from the COVID-19 pandemic and into a new world of hybrid work, organizations will need to intentionally create opportunities for expanded connections that empower workgroups to cultivate innovation.<\/p>\n\n\n\n<h2 id=\"a-note-on-privacy\">A note on privacy<\/h2>\n\n\n\n<p><em>Microsoft takes privacy seriously. We remove all personal and organization-identifying information, such as company name, from our data before analyzing it and creating reports. We never use customer content\u2014such as information within an email, chat, document, or meeting\u2014to produce reports. Our goal is to discover and share broad workplace trends that are anonymized by aggregating the data broadly from those trillions of signals that make up the <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/developer.microsoft.com\/en-us\/graph\/\">Microsoft Graph<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Is innovation another loss due to the global COVID-19 pandemic? Indicators reveal challenges to overcome\u2014as well as opportunities to build on our collective experience gained (opens in new tab) in the last year. Measuring collaboration using network machine learning provides a powerful diagnostic tool to address issues as they arise. One challenge lies in how [&hellip;]<\/p>\n","protected":false},"author":38838,"featured_media":735175,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","msr-author-ordering":[{"type":"user_nicename","value":"Carolyn Buractaon","user_id":"39327"},{"type":"user_nicename","value":"Amber Hoak","user_id":"37992"},{"type":"user_nicename","value":"David Tittsworth","user_id":"38064"},{"type":"user_nicename","value":"Neha Shah","user_id":"40180"},{"type":"user_nicename","value":"Jonathan Larson","user_id":"32385"}],"msr_hide_image_in_river":0,"footnotes":""},"categories":[1],"tags":[],"research-area":[13559],"msr-region":[],"msr-event-type":[],"msr-locale":[268875],"msr-post-option":[243984],"msr-impact-theme":[],"msr-promo-type":[],"msr-podcast-series":[],"class_list":["post-733981","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-research-blog","msr-research-area-social-sciences","msr-locale-en_us","msr-post-option-blog-homepage-featured"],"msr_event_details":{"start":"","end":"","location":""},"podcast_url":"","podcast_episode":"","msr_research_lab":[],"msr_impact_theme":[],"related-publications":[],"related-downloads":[],"related-videos":[],"related-academic-programs":[],"related-groups":[669339,734455,916890],"related-projects":[912408,717493],"related-events":[],"related-researchers":[{"type":"user_nicename","value":"Carolyn Buractaon","user_id":39327,"display_name":"Carolyn Buractaon","author_link":"<a href=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/people\/caburact\/\" aria-label=\"Visit the profile page for Carolyn Buractaon\">Carolyn Buractaon<\/a>","is_active":false,"last_first":"Buractaon, Carolyn","people_section":0,"alias":"caburact"},{"type":"user_nicename","value":"Amber Hoak","user_id":37992,"display_name":"Amber Hoak","author_link":"<a href=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/people\/amhoak\/\" aria-label=\"Visit the profile page for Amber Hoak\">Amber Hoak<\/a>","is_active":false,"last_first":"Hoak, Amber","people_section":0,"alias":"amhoak"},{"type":"user_nicename","value":"David Tittsworth","user_id":38064,"display_name":"David Tittsworth","author_link":"<a href=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/people\/datittsw\/\" aria-label=\"Visit the profile page for David Tittsworth\">David Tittsworth<\/a>","is_active":false,"last_first":"Tittsworth, David","people_section":0,"alias":"datittsw"},{"type":"user_nicename","value":"Neha Shah","user_id":40180,"display_name":"Neha Shah","author_link":"<a href=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/people\/neshah\/\" aria-label=\"Visit the profile page for Neha Shah\">Neha Shah<\/a>","is_active":false,"last_first":"Shah, Neha","people_section":0,"alias":"neshah"},{"type":"user_nicename","value":"Jonathan Larson","user_id":32385,"display_name":"Jonathan Larson","author_link":"<a href=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/people\/jolarso\/\" aria-label=\"Visit the profile page for Jonathan Larson\">Jonathan Larson<\/a>","is_active":false,"last_first":"Larson, Jonathan","people_section":0,"alias":"jolarso"}],"msr_type":"Post","featured_image_thumbnail":"<img width=\"960\" height=\"540\" src=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/1400x788_WTI_still_no_logo-960x540.jpg\" class=\"img-object-cover\" alt=\"\" decoding=\"async\" loading=\"lazy\" srcset=\"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/1400x788_WTI_still_no_logo-960x540.jpg 960w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/1400x788_WTI_still_no_logo-300x169.jpg 300w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/1400x788_WTI_still_no_logo-1024x576.jpg 1024w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/1400x788_WTI_still_no_logo-768x432.jpg 768w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/1400x788_WTI_still_no_logo-1536x865.jpg 1536w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/1400x788_WTI_still_no_logo-2048x1153.jpg 2048w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/1400x788_WTI_still_no_logo-16x9.jpg 16w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/1400x788_WTI_still_no_logo-1066x600.jpg 1066w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/1400x788_WTI_still_no_logo-655x368.jpg 655w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/1400x788_WTI_still_no_logo-343x193.jpg 343w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/1400x788_WTI_still_no_logo-640x360.jpg 640w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/1400x788_WTI_still_no_logo-1280x720.jpg 1280w, https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-content\/uploads\/2021\/03\/1400x788_WTI_still_no_logo-1920x1080.jpg 1920w\" sizes=\"auto, (max-width: 960px) 100vw, 960px\" \/>","byline":"","formattedDate":"March 24, 2021","formattedExcerpt":"Is innovation another loss due to the global COVID-19 pandemic? Indicators reveal challenges to overcome\u2014as well as opportunities to build on our collective experience gained (opens in new tab) in the last year. Measuring collaboration using network machine learning provides a powerful diagnostic tool to&hellip;","locale":{"slug":"en_us","name":"English","native":"","english":"English"},"_links":{"self":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/posts\/733981","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/users\/38838"}],"replies":[{"embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/comments?post=733981"}],"version-history":[{"count":38,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/posts\/733981\/revisions"}],"predecessor-version":[{"id":894498,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/posts\/733981\/revisions\/894498"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/media\/735175"}],"wp:attachment":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/media?parent=733981"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/categories?post=733981"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/tags?post=733981"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=733981"},{"taxonomy":"msr-region","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-region?post=733981"},{"taxonomy":"msr-event-type","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-event-type?post=733981"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=733981"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=733981"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=733981"},{"taxonomy":"msr-promo-type","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-promo-type?post=733981"},{"taxonomy":"msr-podcast-series","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-podcast-series?post=733981"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}