How IIoT deployments go wrong — and what prevents each one.

Industrial IoT projects rarely fail on the sensor or the dashboard. They fail in the middle: on radio coverage nobody surveyed, on integration through a path that was never agreed, and on data nobody owns. Each of these is cheap to prevent and expensive to retrofit.

What goes wrongHow you noticeThe fix
Connectivity chosen before the site surveyCoverage gaps appear after the hardware is installedSurvey first. Concrete, steel, distance and what the building is full of decide the radio — not the datasheet range
Everything streamed to the cloudData volume and cost climb while the useful signal gets harder to findDecide at the edge what is worth sending. Raw telemetry is cheap to generate and expensive to keep forever
No plan for the outageGaps fall exactly across the events you needed to examineBuffer locally and backfill on reconnect — on most South African sites this is a design requirement, not an optional feature
Integration through an undocumented pathReadings stop after somebody makes a control changeRead through a documented interface and agree in writing who may change it and who gets told
Tags named by whoever installed themNobody can find a point six months laterAgree the naming convention before commissioning. Renaming after integration is a migration, not an edit
Clocks not synchronised across sourcesCorrelating two systems produces conclusions that cannot be rightOne clock source for everything. Without it, cause and effect cannot be demonstrated
Thresholds set on go-live dayAlarm fatigue within weeks; alerts acknowledged without being readSet thresholds after a baseline period and review them once more a month later
Security treated as a network question onlyAn exposed controller turns up in an auditSegment the network, expose nothing directly, and stay read-only wherever control is not actually required
No owner for the numbersThe dashboard ages and nobody noticesName the person whose weekly meeting uses the figure. Reports that inform nobody get quietly switched off

Two of these decide whether a deployment survives its second year. The integration path is the first: a reading taken through an undocumented route works perfectly until a controls change breaks it, and by then nobody remembers what it was reading from. Ownership is the second, and it is not a technical problem — a platform with no named consumer for its numbers degrades quietly, because nobody is inconvenienced when it stops being right.