Sensors and dashboards create the impression of progress without measurable improvement.
Critical assets are prioritised, baselines are known, and findings generate work orders.
Too many KPIs, unclear alarms and a conspicuous gap between dashboard and workshop.
Which asset, what failure impact, which intervention, what saving.
Predictive theatre is not fraud – it is the far more common situation where a company has invested in sensors, dashboards and data infrastructure but the decisive step is missing: the connection from data to operational decisions. The sensors deliver values, the dashboard shows curves, but nobody changes a maintenance plan, orders a spare part earlier or reschedules an inspection based on this data.
The pattern is recognisable: a company procures a condition monitoring platform for its fleet. Installation takes months, costs run to six figures. After commissioning, the system generates thousands of data points per day. Technical management receives weekly reports with dozens of KPIs. But the decisive question – which asset should receive which intervention when? – is not answered by the system because thresholds are not calibrated, baselines were never defined or simply nobody is responsible for implementing the recommendations.
The second typical mistake: monitoring too many assets simultaneously. When a system captures 500 data points on a container vessel but only 20 of them are actually action-relevant, the relevant signals drown in noise. The crew or superintendent cannot distinguish between normal variation and a genuine alarm signal. The result: alarms are ignored because too many false alarms occur – a classic boy-who-cried-wolf problem.
The third mistake concerns data quality. Sensors in the maritime environment face harsh conditions: vibration, humidity, saltwater, extreme temperatures. A temperature sensor on a turbocharger that has not been calibrated for a year delivers values – but whether those values reflect the actual condition is questionable. Predictive maintenance based on unreliable data is not prediction but chance.
The fourth and frequently overlooked mistake: missing feedback. When a predictive maintenance recommendation leads to an intervention, the actual finding must be fed back. Only then does the system learn whether its predictions were correct. Without this feedback, prediction quality does not improve – the system remains at the level of its initial calibration.
Predictive theatre ties up resources on multiple levels. Direct costs comprise hardware (sensors, gateways, servers), software (platform licences, cloud services) and personnel (data analysts, IT support). For a mid-size fleet of 15 vessels, annual costs can range between EUR 200,000 and 500,000 – without a single failure being prevented.
Indirect costs are harder to quantify but often higher: the crew loses trust in technological solutions when the system constantly generates alarms that have no consequences. The superintendent spends time analysing reports that yield no actionable insights. Management loses willingness to invest in genuine predictive maintenance solutions because the first investment delivered no results.
Opportunity costs are the most serious item: while predictive theatre absorbs attention and budget, the actual failure risks continue. The turbocharger that genuinely needed early warning is overlooked because the system displays 50 other irrelevant alarms simultaneously.
Genuine predictive maintenance works where four prerequisites are met: critical asset identified, baseline defined, thresholds calibrated and action process established. In practice, vibration monitoring on main engines and turbochargers shows the best results because the physical relationships are well understood and the measurement technology is mature.
Oil analysis is another area with proven benefit. Regular lubricating oil samples analysed for metal particles, viscosity and water content can detect bearing damage, cylinder wear and cooling water leaks early. The advantage: the analytics are standardised, results are unambiguously interpretable and consequences clearly definable.
Thermography on electrical switchgear reveals loose connections and overloaded components before they lead to failures. The investment is modest (a thermal imaging camera and training), the benefit directly measurable. No dashboard required – a trained electrician with a camera and a defined inspection process suffices.
Five questions help distinguish substance from theatre. First: can you name three concrete cases where the system led to a better decision? Second: how many of the alarms generated actually led to an intervention? Third: are the baselines for the monitored assets documented and current? Fourth: is there a defined process for who does what when a warning is issued? Fifth: are findings fed back after an intervention?
If more than two of these questions are answered with no, the probability is high that the system produces theatre rather than substance. The solution is not to scrap the system but to focus it on a few critical assets and backfill the missing processes.
Free initial consultation – we analyze your situation and find the best path forward.
Request Consulting