It is entirely legitimate to choose a reasonable level of security rather than an extreme one. It is equally legitimate to demand exceptional protection when the stakes or the threat justify it. What is inconsistent, however, is to demand that protection while negotiating away, one by one, the constraints that make it effective.
Following the Logic Through
A high level of security is never merely a matter of stronger equipment, more sophisticated technology or a larger number of personnel. It is a system in which the architecture, procedures, access controls, personnel selection, training, drills and daily discipline all meet the same standard. A reinforced door cannot compensate for an uncontrolled exit; advanced technology cannot correct a procedure that nobody follows.
Moving a single piece does not necessarily bring the entire system down, but it changes its balance. Every exception must therefore be treated as a security decision in its own right: it requires either a credible compensating measure, a revision of the intended level of protection, or explicit acceptance of the residual risk. There is no coherent fourth option.
Designing security around the worst-case scenario does not mean claiming that it will occur. It simply means that, once the decision has been taken to protect against it, the system must remain consistent with that assumption. Sometimes the entire contradiction comes to rest on an apparently trivial detail: a window that someone would like to be able to open.
Several Tonnes of Glass — and One Open Window
A true but anonymised account — Certain locations, circumstances, deadlines, quantities and operational details have been altered to preserve the confidentiality of the assignment.
Several years ago, as a major international summit in Asia approached, a highly prestigious hotel was preparing to accommodate a particularly exposed head of state. Several adjoining suites had been reserved, and the requirement from the presidential security team was unequivocal: every glazed surface had to provide a very high level of ballistic protection within an exceptionally short timeframe.
Following an immediate feasibility assessment, the hotel accepted our working conditions and the assignment could begin. The selected solution consisted of fixed BR7-rated glazing, 72 millimetres thick in this configuration. At that level, the glass, frame, fittings, anchors and their integration into the building must function as a single system. A comparable opening unit would have required a bespoke design that could not be delivered within the available time.
Work was progressing rapidly when, midway through the installation, a new request was made: one of the windows now had to open. The reason was entirely human: the head of state liked being able to open a window and enjoy the outside air. In an ordinary hotel room, such a preference would have attracted no comment. But this was no longer an ordinary hotel room.
The presidential team, fully aware of the risks and threats, had selected protection intended to address a prepared attack conducted from a distance. That choice did not determine the range of a potential shot or the precise profile of an assailant. It did, however, require us to take seriously the possibility of an operation carried out by someone trained, equipped and capable of studying the location, angles, materials and openings. In such a scenario, an ordinary window was no longer a minor concession: it became the means of bypassing everything else.
The difficulty was compounded by the window’s exposure to long sight lines that formed part of the threat assessment. To a trained observer, differences in tint, reflection or construction could make the opening easier to identify. Several tonnes of highly resistant glazing could not compensate for a single weak point placed precisely where it had been requested.
We therefore formally advised against the modification. The discussion continued, caught between the protection requirement, the deadline and a personal preference that the team, under considerable pressure, found difficult to refuse. I eventually summed up the situation with an image: “You do not fit bulletproof glass to a canvas tent.” My counterpart smiled. The expression was deliberately exaggerated; the reasoning was not.
The request was nevertheless maintained. We required a written instruction clearly recording our reservation, the recommendation made and the acceptance of the resulting residual risk. That record plainly did not turn the window into a protected opening or erase anyone’s responsibilities; it established the advice that had been given and who had taken the decision. In the interests of completing the work properly, we nevertheless implemented a compensating measure intended to make the difference in appearance less readily identifiable, without claiming to restore the ballistic performance that had been relinquished.
The assignment was completed within the deadline. The visit passed without incident; the temporary installation was then removed and the original glazing reinstated. Everything had gone well. But the absence of an incident did not validate the exception, any more than it proved the value of the tonnes of glass that had been installed. It meant only that the anticipated scenario had not occurred.
The Client Sets the Objective; the Professional Directs the Operation
This story extends far beyond glazing. The higher the assessed threat level, the more fundamentally the relationship between client and professional changes. The client defines the purpose, constraints and budget, and approves the overall framework. But engaging an expert is not a way of purchasing their approval; it is a decision to entrust them with the technical coherence of the measures required to achieve the stated objective.
Close protection offers another example. A request for discreet accompaniment allows for a relatively flexible arrangement. Requiring an armed Close Protection Officer (CPO), by contrast, introduces the assumption of a serious and targeted threat: the assignment is no longer a convenience service to which a weapon has simply been added. The client retains control of the purpose of the journey, while the protection team assumes operational direction.
When a client rejects a recommendation, the consultant must explain the consequences, identify possible compensating measures and formally record the residual risk that has been accepted. Such traceability clarifies responsibility; it does not make the decision safer. If the request becomes unlawful, technically indefensible or incompatible with professional obligations, the professional must also be prepared to refuse it.
The value of an expert therefore lies not only in the ability to complete a difficult assignment. It also lies in their independence: knowing what is possible, what is not, and the precise point at which a concession ceases to be an adjustment and becomes a breach in the system.
Security is, ultimately, a commitment to consistency. Anyone asking to be protected against the worst must accept a system designed as though the worst could genuinely occur. Otherwise, we are no longer building protection; we are merely maintaining a sense of security. And once again, we end up fitting bulletproof glass to a canvas tent.