No structural monitoring has no line item in the budget, so it looks free. You pay for it elsewhere: in "just in case" site visits, in expert reports ordered from scratch, in downtime caused by uncertainty, in repairs postponed until they are more expensive, and in disputes lost for lack of data. Below are seven such costs, formulas to calculate them on your own, and a payback model - without figures and without outside statistics.
In brief
- The cost of no monitoring is the sum of the losses you incur because you learn about a change in the condition of the asset too late, incompletely, or not at all.
- Three costs are certain and you pay them every year: excess inspections, an asset maintenance budget divided blindly, knowledge of the asset that leaves with people.
- Four are expected (probability × impact): downtime caused by uncertainty, a delayed decision, a weak position in a dispute, reputation.
- Each can be approximated with a simple formula using your own numbers. This text contains not a single amount - intentionally.
- Monitoring starts to pay back when the data-limited part of the certain and expected costs exceeds the annual cost of the system. At the end there is a table to fill in.
Why "how much does the system cost" is the wrong question to ask first
In a discussion about structural monitoring, the first question is almost always: how much does the system cost? The question makes sense, but only as the second one. The first should be: how much does it cost me that today I know about the condition of the asset only as much as the last inspection or the last surveyor reading showed?
The cost of no structural monitoring is the sum of the losses an organisation bears because information about a change in the asset's condition reaches it late, in incomplete form, or not at all. Structural health monitoring (SHM) is, in turn, the continuous measurement of selected physical quantities - inclinations, strains, deflections, vibrations, temperature - with thresholds, alarms and a history that makes it possible to distinguish normal asset behaviour from a trend that requires action. A broader introduction is available in the complete SHM guide.
The difference between those two numbers is what should go before the board. One is on the quote. Nobody has written down the other one - and that is why it loses every comparison, although it is usually larger. Below, we break it down component by component: first the costs already in your invoices, then the ones that appear only when something happens.
Two caveats from the outset. All scenarios in this text are hypothetical - they do not describe any specific asset or client. We present formulas as fields to be filled in, because only your own numbers make sense for your own asset; someone else's failure statistics tell you nothing about your own hall.
Certain costs, part 1: inspections and measurements "just in case"
Start with the certain costs. They do not require probability estimates - they are in invoices, travel claims and timesheets. If you calculate only one thing from this text, calculate these.
1. Excess inspections and measurements "just in case"
Without continuous data, an organisation compensates for uncertainty with more frequent site visits. Every visit means travel, labour hours, access coordination, a report. And each one shows the condition at a single point in time, not a trend.
Scenario (hypothetical): after a severe winter, the manager of a warehouse hall sends an engineer to inspect the girders after every significant snowfall and after every tenant phone call. In one season this adds up to a dozen trips; none of them reveals anything worrying. The money is spent, but there is still no trend: nobody knows whether the deflection returns to zero after thawing or whether a little more remains every year.
Some of those trips are a duty, not excess. Article 62(1) of the Polish Construction Law requires periodic inspections at least once a year and at least once every 5 years, and for buildings with a built-up area over 2 000 m² and other structures with a roof area over 1 000 m², at least twice a year, by 31 May and by 30 November. In addition, there is an inspection of safe use after external factors, for example heavy rainfall, have occurred (Article 62(1)(4) in conjunction with Article 61(2)). Those inspections remain. The excess cost is everything in between: trips driven by uncertainty, because nobody knows whether the roof is behaving normally.
How to calculate it on your own:
- cost of one inspection = travel + labour hours (including preparation and report) + equipment + access cost (MEWP, area closure);
- number of inspections per year that you do because of uncertainty, not because of a legal requirement or schedule;
- annual cost of excess inspections = cost of one × number of "uncertainty-driven" visits.
Monitoring changes extra inspections from a weather reaction into a data reaction: you send people where the chart shows something, and you carry out the mandatory inspection with a history of deflections in hand, not from scratch.
Certain costs, part 2: blind budgeting and knowledge that leaves with people
Two more certain costs rarely land in the same line item as monitoring, although both come from the same absence: there is no trend, only a snapshot.
2. A maintenance budget divided blindly
In a portfolio of a dozen assets, the problem is not knowing that something needs attention. It is knowing what needs attention first. Without data, the budget is split according to the loudest ticket, the oldest asset, or the most persuasive manager.
Scenario (hypothetical): a municipality has a dozen bridges and footbridges. The repair budget covers two assets per year. It chooses the ones with the worst score in the inspection. Meanwhile, a third asset - with a better score - has a deflection trend that has been rising for three seasons. Nobody sees it, because an inspection is a snapshot, not a film.
How to calculate it on your own:
- annual maintenance budget for the portfolio;
- share of expenditure that, in hindsight, turned out to be misdirected - too early, too late, or on the wrong asset (your estimate; for illustration: one fifth, one third);
- cost of misallocation = budget × that share.
You do not need to monitor the whole portfolio to reduce this cost. A trend instead of a single annual snapshot is enough for the assets about which you have the most uncertainty - and how to identify them is described in the post How to Choose an Asset for a Monitoring Pilot.
3. Lost knowledge about the asset
If knowledge about the asset sits in the heads of two people, in folders on a network drive and in spreadsheets named "measurements_final_v3", every team change is a risk. The new manager begins with document archaeology and for the first few months does not know what is normal for that asset.
Scenario (hypothetical): the maintenance engineer who has "felt" the stadium for ten years retires. His successor receives binders and a few spreadsheets. In the first winter, during heavy snowfall, he does not know whether the roof girder deflection is usual or not - because there is nothing to compare it with. He orders an expert report.
How to calculate it on your own:
- cost of rebuilding knowledge = expert reports and measurements ordered "from scratch" after a team change + time needed to onboard a new person;
- annual cost = cost of rebuilding × turnover frequency in maintenance roles in your organisation.
A continuous history of measurements, events and decisions in one place does not replace experience. It gives the successor a reference point from day one - and that is the whole difference between "I do not know whether this is normal" and "last year, in similar snow, it was the same".
Expected costs, part 1: when the asset stops or a repair becomes more expensive
The second group consists of costs that appear only when something happens. You do not pay them every year, but you carry their risk every year - and risk has a price: probability × impact.
4. Unplanned downtime caused by uncertainty
It is easiest to calculate the people and equipment that are standing idle. It is harder to calculate lost traffic, contractual penalties for delay, a shifted handover, and the fact that the decision "do we stop or keep working" is made on caution, not on data.
Scenario (hypothetical): a contract manager is carrying out excavation next to an existing building. On Monday morning, the caretaker of the neighbouring tenement reports a crack on the staircase. Survey measurements are taken once a week, the last one was on Thursday, so the manager does not know whether the crack is new or has been there for years, or whether the slurry wall moved over the weekend. He stops the works until the surveyor and an expert can attend. The excavator, crane and three crews wait for two days. With a continuous record of building inclinations and wall displacements from the last 72 hours, the decision could have been made that same morning - either way.
How to calculate it on your own:
- cost of one day of downtime = daily labour cost + daily equipment cost + daily contractual penalty or lost revenue (enter your own rates);
- number of downtime days "due to uncertainty" = how many days pass from the report to a reliable assessment under the current measurement method;
- annual cost of downtime caused by uncertainty = cost per day × days per event × number of such events per year (estimate from your own portfolio).
Monitoring does not eliminate risk. It shortens the time from the question "can we keep working?" to an answer based on data - and that time is exactly what costs money. Eurocode 7 (PN-EN 1997-1, observational method) states this directly: the response time of instruments and analysis procedures should be sufficiently short in relation to the possible evolution of the system. How to calculate that time against periodic measurement is described in the post “100× faster” - from event to decision.
5. A decision made too late
Serious damage rarely appears without warning signs. The signs are there, but they are scattered, irregular and noticed after the fact. In infrastructure, a late reaction almost always means a more expensive repair: larger scope, emergency mode, closure of the asset instead of work in a planned window.
Scenario (hypothetical): a road authority manages a viaduct with a bearing that is slowly losing its ability to move freely. During the annual inspection, the inspector notes "to be observed". A year later, the bearing is jammed, the expansion joint is functioning improperly, and cracking appears in the deck. A planned repair on one lane becomes a refurbishment with asset closure. Continuous measurement of bearing displacement against temperature would have shown that the bearing had stopped keeping up with daily temperature changes - months before the next inspection.
How to calculate it on your own:
- cost of planned repair (contractor estimate in normal mode);
- cost of emergency repair of the same element (urgent mode + larger scope + asset closure);
- difference = delay premium; multiply it by your own estimate of how often "to be observed" becomes "urgent" in your portfolio.
Monitoring shifts the conversation from "what happened?" to "what is starting to change?". For you, that means choosing the timing and scope of the repair instead of accepting the timing imposed by failure.
Expected costs, part 2: disputes and reputation
The last two costs are the hardest to calculate and the easiest to dismiss - until the first claim.
6. A weak position in a dispute
When a claim appears - from a construction neighbour, investor, insurer or subcontractor - an opinion without data has limited force. A time-stamped measurement series, an alarm log, and a record of who reacted and when show that decisions were taken with due care.
Scenario (hypothetical): the owner of a building next to a construction site claims that cracks appeared during piling works. The contractor has a zero-condition survey from a year ago and nothing else. The dispute comes down to expert opinion versus expert opinion. If the contractor had a continuous record of building vibrations from the piling period - assessed in line with PN-B-02170:2016-12, meaning the peak values of horizontal components in one-third-octave bands measured at foundation level - and a record of inclinations, the discussion would look different, regardless of where the truth lies.
How to calculate it on your own:
- cost of handling one dispute = lawyers + experts + management labour hours + any settlement or compensation;
- estimated difference in outcome between a dispute "with data" and "without data" - there is no formula here, only your legal counsel's assessment;
- annual cost = probability of a dispute in your portfolio × difference in outcome.
There are two bases of liability, and both favour data. Article 415 of the Civil Code links liability for damage to fault - and it is easier to prove the absence of fault by showing what was measured and when. Article 435(1) of the Civil Code goes further: a business operated with the aid of natural forces (courts include construction businesses in this category) is liable for damage regardless of fault, and is released only by force majeure or the exclusive fault of the injured party or a third party - in that case, data are used to show where the damage really came from. On the side of the owner or manager of the asset, Article 61 of the Construction Law also applies: the duty to maintain the asset in proper technical condition and to ensure safe use with due care. What monitoring data actually work as evidence is unpacked in the post Monitoring Data as Evidence in a Dispute.
7. Reputation cost
Asset closure, an incident, a media conflict around safety - this does not end with a cost estimate. For a contractor bidding for tenders, procedural reputation (we had data, we had a procedure, we reacted) can be as important as price.
Scenario (hypothetical): a general contractor loses a point in tender evaluation because the references from a previous contract mention work being halted by the building supervision authority after complaints from neighbours. Nobody investigated whether the risk was real. It was enough that the contractor could not show data at the time.
How to calculate it on your own:
- value of the contracts you bid for each year × your estimate of how often procedural reputation decides the result;
- or, more simply: how many times in recent years you had to explain an incident without data and how many management labour hours it consumed.
You will not defend this cost to the board as a number. You will defend it as a question: next time, do we want to explain ourselves with data or without them.
The 7 costs in one table: who pays and where to find them in the budget
| # | Cost | Who pays for it | Where to find it in the budget |
|---|---|---|---|
| 1 | Excess inspections | asset manager, maintenance department | outsourced services, travel, engineers' labour hours |
| 2 | Budget divided blindly | portfolio owner | maintenance budget - shifted or repeated line items |
| 3 | Lost knowledge | maintenance department | expert reports "from scratch", onboarding time for new staff |
| 4 | Unplanned downtime | contractor, asset manager, users of the asset | indirect contract costs, contractual penalties, lost revenue |
| 5 | Late decision | asset owner | emergency repairs instead of planned ones, contingency reserve |
| 6 | Weak position in a dispute | contractor, investor, insurer | legal services, compensation, settlements, claim reserves |
| 7 | Reputation | management, tendering department | lost contracts, management time, crisis communications |
These costs are already in your books. They are simply scattered across items that nobody connects with the word "monitoring". You can fill in the first three rows from invoices; for the remaining ones, you need a probability assessment - and that is what the next section is about.
When monitoring starts to pay back: a one-page model
A simple model that fits on one page. It divides the costs of missing data into the two groups from the previous sections: certain costs (you incur them every year) and expected costs (probability × impact).
Step 1. Annual certain costs (A). Excess inspections (cost 1) + misallocated budget cost (cost 2) + annualised knowledge rebuilding cost (cost 3) + people's time spent manually collecting data from loggers and preparing reports.
Step 2. Annual expected costs (B). For each of costs 4-7: probability of the event in a year × impact of the event. Add them up. Be careful - an understated estimate you can defend is worth more than a polished one that someone can challenge with the first question.
Step 3. Annual monitoring cost (C). Full TCO (total cost of ownership) on an annual basis: sensors and installation spread over the service life, platform, transmission, maintenance, calibration. Component by component - in the post What Structural Monitoring Cost Consists Of (TCO).
Step 4. Reduction factor (r). Monitoring does not eliminate costs A and B. It reduces them. Enter your own cautious estimate of what share of each cost will be realistically reduced by continuous information with thresholds and alarms. For downtime caused by uncertainty this may be a lot, for reputation less.
Result: monitoring starts to pay back when (A + B) × r > C.
For illustration: let us assume that an event forcing two days of downtime because of uncertainty occurs once every five years (probability 0.2 per year), that you have two excess inspections per year, and that continuous data reduce both costs by half. No amounts are given here - enter your own rates from steps 1-3 and see on which side of the inequality your asset lands. In practice, it is often not downtime that decides, but the boring, certain item A.
| Field | Where to get the number | Your value |
|---|---|---|
| A - certain costs/year | invoices for inspections, labour hours, expert reports | ... |
| B - expected costs/year | probability × impact for costs 4-7 | ... |
| C - annual TCO of monitoring | quote + TCO item | ... |
| r - share of costs reduced by data | your cautious estimate | ... |
| (A + B) × r - C | if positive: monitoring pays back | ... |
It is worth taking that page to a board meeting. How to defend it and what questions to expect is covered in the post How to Justify Structural Monitoring to the Board.
Three objections you will hear - and what to do about them
"We will overestimate it and end up looking alarmist." That is a real risk - which is why the model starts with certain costs. If item A alone exceeds C, you do not need to discuss failure probabilities at all. If it does not exceed C, only then do you move to B, carefully and using your own numbers.
"False alarms will cost more than missing data." Poorly set thresholds do create a cost: trips for nothing and the habit of ignoring notifications. That is why thresholds are set with hysteresis, based on the history of the specific channel, not from a catalogue, and missing data are treated separately (NO_DATA state) as a signal about the equipment, not the structure. The first season with data is used to learn how the asset "breathes" daily and seasonally; that must be honestly built into the schedule.
"It is not mandatory, so why bother?" True: regulations require periodic inspections, not continuous monitoring (exceptions follow from the design, supervision decision, contract or policy). But all seven costs on this list arise between inspections, in periods the regulation says nothing about. The regulation sets the minimum. The cost of missing data is set by your wallet.
What this looks like in Inclify
Inclify is an online platform that turns the costs on this list into data. On the asset, sensors from a team with 15 years of experience in measurement systems "from stadiums to hydraulic locks" work - vibrating wire, inclinometers, strain gauges, MEMS, piezometers, temperature sensors - or your existing loggers, connected through HTTP/JSON; by default, they send readings every 15 minutes by default, or more often. Each channel has a warning threshold and an alarm threshold with hysteresis; when one is exceeded, you receive SMS and e-mail, and you confirm the alarm in the panel (with a record of who took over and when). A chart with two axes - for example, inclination against temperature - helps distinguish daily "breathing" from a lasting trend before you send someone to site.
The measurement history in UTC, the alarm history and the audit log of configuration changes remain in the database without automatic deletion; you export the data table to CSV when an expert or successor needs it. The data SLA report shows completeness and gaps, so you do not have to check manually whether the loggers are alive. You have the entire portfolio of assets in one panel. If the asset already has sensors, connection takes a few days; deployment with installation takes from a few to a dozen weeks. More on the monitoring platform page; quote on request.
FAQ
Is no structural monitoring a breach of regulations?
As a rule, no. Article 61 of the Construction Law imposes on the owner or manager the duty to keep the asset in proper technical condition and ensure safe use in the presence of external factors, while Article 62 requires periodic inspections. Neither directly requires continuous monitoring. Monitoring may be required by the design, a supervisory decision, contract terms or an insurance policy. This text is not legal advice - for a specific case, consult a lawyer.
How do you calculate the cost of no monitoring if there is no failure data?
Start with certain costs: extra inspections, labour hours for manual data collection, expert reports ordered from scratch. Those numbers are in invoices. Only then estimate expected costs (probability × impact) - cautiously, from your own portfolio, not from somebody else's statistics. A cautious estimate that you can defend is worth more than a striking one that nobody can verify.
Does monitoring eliminate downtime?
No. It eliminates downtime caused by uncertainty - when the asset stands still because nobody knows whether work can continue. If the data show a real hazard, the asset still has to be shut down, only earlier and with justification. Monitoring shortens the time from question to answer; the engineer makes the decision.
Does this make sense for one small asset, not a portfolio?
It depends on what uncertainty costs you. One asset where downtime, a dispute or a work stoppage really hurts can justify monitoring on its own. If you are not sure, a sensible starting point is a pilot on one asset and a comparison of certain costs "before" and "after" over one season.
How is the cost of no monitoring different from the cost of failure?
The cost of failure is one item within the cost of no monitoring - and often not the largest one. Most organisations pay mainly for uncertainty: extra trips, decisions delayed until the next inspection, expert reports after a team change, a weaker position in disputes. Those costs appear every year, without any failure.
How quickly will monitoring start reducing these costs?
Certain costs - trips driven by uncertainty, manual data collection - begin to fall from the first weeks with data, when you see a trend instead of a single snapshot. Expected costs need a reference point: usually one season to learn the daily and seasonal behaviour of the asset and set thresholds that do not create false alarms for no reason. Connection of existing sensors takes a few days; deployment with installation takes from a few to a dozen weeks.
Sources and further reading
- Act of 7 July 1994 - Construction Law, Articles 61 and 62 (duties of the owner and manager, periodic inspections) - Internet System of Legal Acts, isap.sejm.gov.pl (legal status as of August 2026).
- Act of 23 April 1964 - Civil Code, Articles 415 and 435 (liability based on fault and on risk) - isap.sejm.gov.pl.
- Judgment of the Supreme Court of 17 March 2022, II CSKP 482/22 (general contractor as an enterprise set in motion by the forces of nature; damage to a neighbouring building by construction works) - sn.pl.
- PN-B-02170:2016-12 Assessment of Harmfulness of Vibrations Transmitted through the Ground to Buildings (SWD scales) - Polish Committee for Standardization, sklep.pkn.pl.
- PN-EN 1997-1 Eurocode 7: Geotechnical Design - Part 1: General Rules (observational method: monitoring plan, instrument response time, contingency plan) - sklep.pkn.pl.
- Inclify posts: What Structural Monitoring Cost Consists Of (TCO), How to Justify Structural Monitoring to the Board, Monitoring Data as Evidence in a Dispute.
What next
Fill in the table from the section "When monitoring starts to pay back" - first item A only, from invoices. If the result is higher than you expected, or you do not know how to estimate C for your asset, let's talk. We reply within 24 hours and show how it looks on an asset similar to yours - without obligation. The lowest entry point is a pilot on one asset - the one where uncertainty costs you the most - launched before the season in which the asset works hardest: before winter for a hall, before the excavation phase for a construction project, before the next periodic inspection for a manager. If the asset already has sensors or loggers, connection takes a few days and the first season with data starts without installation. Quote on request.