What is Shadowing?

Shadowing is an observational research method in which the researcher accompanies a participant through a period of their normal activity (a working day, a shift, a specific process from start to finish), observing continuously and asking questions as the activity unfolds. It sits in the family of ethnographic and observational methods, distinguished by its focus on following one person's trajectory rather than watching a place, and by its light-touch questioning: the researcher is present and curious but the participant leads. Its closest relative is contextual inquiry, formalised by Hugh Beyer and Karen Holtzblatt, which frames the researcher as an apprentice learning the work from the participant as master; shadowing shares the setting and the humility and typically runs longer and more passively.

What Shadowing Reveals

The gap between described and actual work. Interviewees describe the ideal process; shadowing shows the spreadsheet on the second monitor, the sticky note with the password, the six-tab workflow that replaced the official tool. These are the jobs and pain points people don't report because they've stopped noticing them.

Context and interruption. Software is designed for focused users and used by interrupted ones: the phone call mid-form, the colleague's question, the ten-minute gap before the task resumes. Shadowing captures the rhythm of real use in a way no session can.

Tacit knowledge. The judgment calls experts make without being able to articulate them, visible only when they're made in front of you, and askable in the moment ("what made you check that?").

The ecosystem. Who else is involved, what artefacts flow between them, where handoffs fail: the material of journey maps built from evidence rather than assumption.

Doing It Well

1. Negotiate access and consent carefully.
Shadowing exposes a person's whole working context, including colleagues who didn't sign up; consent covers the participant, and discretion (pausing observation, not recording bystanders) covers everyone else.

2. Run long enough for normality to return.
The observer effect is strongest in the first hour; discount the early performance and let the day settle.

3. Observe first, ask second.
Questions interrupt the very behaviour you came to see; batch them for natural pauses, and prefer "what just happened there?" to hypotheticals.

4. Record thickly and immediately.
Timestamped field notes, photos of artefacts and environments (with permission), and a debrief written within hours, while the detail is intact; the raw material of later analysis.

5. Shadow for range.
A few participants across roles, sites, and expertise levels reveal the variation a single day can't; sampling is purposive and should be planned as such.

Scaling the Insight

Shadowing is expensive per participant and unmatched per hour of insight, which argues for using it early, on a few people, to discover the questions that cheaper methods then answer at scale: a workaround seen in two shadowing days becomes a survey item for a thousand users, or a prototype task for a recorded remote study. Where physical presence is impossible, lightweight remote variants (screen-share sessions, participant-recorded walkthroughs of their own workflow, the kind of self-captured video a Ballpark study collects) trade the researcher's eye for the participant's camera and recover a useful fraction of the context.

The Takeaway

Shadowing follows the work where it happens, watching the actual process, the interruptions, the tacit judgments, and the ecosystem that interviews can't reach. Secure consent broadly, wait out the performance, observe before asking, record thickly, sample for range, and use what you find to aim the scalable methods. The process document says what should happen; a day at someone's shoulder says what does.