Scientific research is increasingly conducted through distributed, digitally mediated collaboration. Even relatively small scientific teams now do much of their work through shared platforms that support communication, data sharing, document development, and workflow coordination. These platforms do more than enable work. They also generate persistent behavioral records of how teams communicate, coordinate, and produce scientific outputs. We refer to these records as digital exhaust and argue that they provide an underused opportunity to study teamwork and knowledge building as they unfold in digitally mediated environments. Although the Science of Team Science has generated important insights into interdisciplinary collaboration, research on scientific teamwork still relies heavily on surveys, interviews, and episodic observation. These approaches provide important perspectives, but often capture only limited slices of interaction and output. Digital trace data can complement these methods by providing more continuous indicators of collaborative behavior embedded in the infrastructure of scientific work.

This presentation introduces a socio-technical measurement framework for analyzing digital exhaust generated through collaborative research environments. Rather than viewing collaboration technologies only as tools for supporting work, we conceptualize them as socio-technical infrastructures that both scaffold and record the cognitive and organizational processes of teams. Our framework focuses on three aspects of socio-technical organization: the structuring of people, work, and events. The organization of people includes role assignments, communication channels, and documentation practices that support coordination. The organization of work includes protocols, collaboration agreements, and version-controlled workflows that shape how artifacts are created and revised. The organization of events includes recurring meetings and milestone-driven activities that structure collective progress. Together, these scaffolds function as cognitive artifacts that help teams externalize and maintain shared understanding.

Using this perspective, we outline how digital traces associated with collaborative artifacts, such as document revision histories, repository logs, meeting records, and communication patterns, can be mapped onto constructs studied in team cognition and the Science of Team Science. These traces can illuminate how teams manage shared knowledge over time, including how artifacts accumulate, diverge, and are reintegrated across asynchronous work. This is especially important in distributed contexts, where shared understanding may degrade through fragmentation or uneven access to evolving knowledge. Trace data can therefore support not only retrospective measurement, but also the design of AI-enabled teamwork assistants that monitor knowledge artifacts, flag divergence, trace lineage across revisions, and support re-synchronization at key collaborative transitions.

We argue that analyzing digital exhaust within a theoretically grounded framework extends the methodological toolkit for studying scientific teamwork while identifying practical entry points for intelligent support. Rather than replacing established qualitative approaches, trace-based measures can augment them by providing fine-grained, longitudinal views of collaboration as it occurs in practice. More broadly, this approach supports the development of a collaboration observatory capable of integrating behavioral traces across platforms to generate interpretable indicators of participation, coordination, knowledge co-construction, and team health in digitally mediated scientific teams.