Technology probes collect usage data in situ
Aliases: in-situ technology probe · field-deployed simple technology
What it is
A technology probe is a simple technical object placed in real life: usable, instrumented, deliberately not a finished product. Hutchinson and colleagues introduced it to collect actual use traces in homes and similar settings, and to let design ideas collide with use, rather than to demo a feature list in a lab. It differs from a usability prototype in place and purpose: the probe asks what people do with it in their own environment; the test prototype asks whether a prescribed task can be completed. It also differs from design probes discussed alongside it—those elicit expression; this one records use.
Why it happens
Lab tasks strip out household rhythm, visitors, parallel chores, and shared devices, which are the real constraints on domestic and field systems. Leave a sparsely featured, fully logged device in place, and use unfolds on life’s timetable: who touches it, ignores it, modifies it, or puts it to a job the designers never imagined. Simplicity makes behavior attributable—too many features and log patterns cannot be traced to an item. Logging captures frequency and timing that self-report misses. The field also changes social relations: the device becomes a new household member, and negotiation, avoidance, and display enter the data.
Studying it
Typical deployments last weeks, in a few homes or workplaces, with the log scope and right to withdraw stated up front. Hutchinson and colleagues treated the technology probe as social research, engineering trial, and design inspiration at once, so analysis runs three lines: temporal patterns in logs, technical failures and modifications, and new wishes in interviews. Log fields must be aligned to questions in advance (who, when, which action) or they cannot be counted later. Interviews should follow a stretch of unprompted use so first-day novelty is not treated as stable practice. Do not treat the probe as an A/B test: sample, controls, and feature completeness cannot support an effect estimate.
Where it stops holding
A “real environment” is still one entered by consent, not an unbiased natural history. Households willing to take a device home are already skewed. Safety, privacy, and bandwidth forbid some logs. In public space or high-risk settings, an unattended technology probe may be unacceptable. If the goal is to check whether an already-chosen interface is usable, the probe is too heavy and too wide; use a task test. Children, care settings, and shared housing need extra consent and observation bounds.
Applying it
- Cut features to the minimum that can answer the field question, and write log fields into the ethics notice.
- Recruit people who actually live in that environment, not lab-convenient participants who “take it home for a night.”
- Visit after deployment, read logs before asking why, and do not fill gaps with recall alone.
- Write outputs as use patterns and unexpected jobs, not as a requirements list or a success rate.