How These Deployments Fail

Location deployments fail in recognisable ways, and the diagnosis matters more than any implementation plan.

Tracking never stops

The symptom: records exist for Saturday evening.

The cause: continuous collection is the default and suspension requires configuration.

The correction: automatic suspension outside shifts, a working pause control, or event-based collection that has no out-of-hours dimension.

Collecting a trail to answer a point question

The symptom: continuous position stored to prove attendance.

The correction: geofence events, and turn the trail off. The events answer the question better and collect a fraction as much.

Dwell time becomes a performance measure

The symptom: rushed jobs, rising callbacks, difficult work avoided.

The correction: aggregate by job type for planning; never per person for comparison.

Retention left at the default

The symptom: six months of positions, and a subject access request that takes a week.

The correction: a period per category, automated deletion, verified including backups and exports.

No pause control, or one nobody uses

The symptom: personal journeys in the record.

The correction: enable it, demonstrate it, and check it is being used. An available feature that was never switched on has been treated as an aggravating factor by regulators.

Gaps treated as evasion

The symptom: a disciplinary process that collapses when the car park is mentioned.

The correction: a dead zone map, device health monitoring, and a stated rule that gaps are questions.

Consultation skipped

The symptom: a works council objection after procurement, or a workforce that learned about it from the installer.

The correction: consult before buying, respond in writing, publish what changed.

No owner afterwards

The symptom: geofences that drift out of date, access never reviewed, retention never enforced.

The correction: a named person with allocated time.

Cameras arriving as a free module

The symptom: driver-facing recording nobody decided to deploy.

The correction: decide each direction separately, record the decision including the one not to enable.

The common thread

Each is a default left in place rather than a choice made.

The products in this category are configured to collect everything, and the entire discipline of deploying one well consists of turning things down and writing down why.

The recovery move

For a deployment already generating complaints.

Publish what is actually collected, including what is not.

Turn off out-of-hours collection and say so.

Enable and demonstrate the pause.

Shorten the retention and verify the deletion ran.

Report one finding about the work and fix it visibly.

One quarter of work, and it converts a resented system into one people leave alone.

Connect policy to configuration

The practical choices behind this note can be compared with this policy example. Keep the written purpose in control of the setup, enable only the data needed and review the result with affected workers.