The major crises an organisation goes through were almost always signalled internally before they happened. Operational people saw them, managers escalated them, abnormal data was produced. And nothing moved.
That observation has become so familiar it has lost its force. It deserves to be taken seriously, because it points to a precise problem. The problem is almost never the ability to perceive the weak signals, it is the architecture that should let them reach a decision.
Curiosity for weak signals is the third of the eight pillars of the invulnerable company. It is the capacity to capture, escalate and process information that is available but under-interpreted, what I call trendability, before the threat becomes obvious.
| The observation | The mechanism | The sequence |
|---|---|---|
| The information already existedCuriosity for weak signals starts from a fact leaders have lived without naming: the major crises had been signalled internally before they struck. The signals rarely go missing, the institutional paths that would carry them to a decision in time are what is absent, and nobody builds them while nothing is happening. | Fear filters before the hierarchy doesResearch shows the filtering is not deliberate. Middle managers fear their peers and superiors, which lowers their propensity to pass negative information upward, and produces at the top an over-optimistic reading of the organisation’s real capabilities. | Capture, process, decide, recordInstalling this pillar follows an order. Channels independent of the reporting line, a body dedicated to processing, an explicit authority to trigger a decision, and a written trace of the signals set aside with the reason they were. |
Why the signals exist and never arrive
The temptation, when analysing major organisational crises, is to look for the culprits. The leader who ignored the alerts, the team that hid the problems, the board that failed to ask the right questions. Those explanations are not wrong, they are incomplete, and they hide a more useful reality.
What a weak signal is
A weak signal is available but under-interpreted information. It exists in the organisation before the threat is visible, sometimes years before, and it is carried by individuals who perceive it, most often in the operational teams.
That definition excludes classic competitive intelligence. When a signal is strong enough to appear in a sector report, it is no longer weak. It has already had time to harden into a constraint, and the competitive advantage it carried has gone.
It also excludes visionary individuals. They exist in every organisation, including the ones that collapse. Their presence has never been enough, and relying on them means entrusting the survival of the organisation to the luck of having hired the right person in the right place.
The problem is processing, not perception
The difficulty sits between individual perception and collective decision. A signal captured by an operator has to cross several hierarchical levels, each with its own interests, its own priorities and its own reading of what is worth passing on.
At each crossing, the signal can be softened, reworded or held back. That filtering is no conspiracy, it is the rational behaviour of any hierarchical organisation that has learned bad news costs the person who carries it.
This is why the pillar rests on mechanisms rather than on an open culture or participative management. Those elements create favourable conditions without constituting the apparatus, and with no formal process guaranteeing the signal reaches the decision, an open culture stays a promise with no infrastructure.
What the research establishes about the filters
The filtering mechanism has been documented precisely, and the result contradicts the usual explanation by complacency or incompetence. It points to a cause that is far harder to correct, because the people doing the filtering are behaving rationally.
Boeing, five years of signals with no architecture to receive them
Boeing engineers raised serious concerns, repeatedly and through official internal channels, about the certification procedures for the flight control system of the 737 MAX. Those signals go back several years before the first accidents.
They were not ignored out of malice. They were absorbed by an organisational architecture that was not designed to process them. The financial performance culture installed from the late 1990s had progressively narrowed the space in which contradictory technical signals could weigh on commercial decisions.
The escalation mechanisms existed on paper. They no longer had the conditions to work in practice, which is a very different situation from their absence, and a much more dangerous one, because it produces a sense of vigilance without producing its effects.
Kodak, the signal that threatens the model carrying it
The Kodak case is different in form and identical in logic. An engineer at the company invented the first digital camera in 1975, and the group was filing patents on those technologies through the 1980s. The absence of signals was therefore never the problem.
The structural incapacity was about turning those signals into decisions that questioned the dominant model. Film generated considerable margins, and every signal pointing toward digital directly threatened that model.
Kodak’s architecture, its investment appraisal processes and its performance culture were calibrated to maximise the value of the existing business. They were not built to hear what questioned it, which is the most expensive blind spot an organisation can construct.
Field note
Nokia, the asymmetric fear the research measured
In 2007 Nokia held more than 40 per cent of the global mobile phone market. Six years later its phone division was sold for a fraction of its value. Two researchers, Timo Vuori of Aalto University and Quy Huy of INSEAD, conducted seventy-six interviews with executives, middle managers, engineers and external experts to understand the fall between 2005 and 2010.
Their conclusion, published in Administrative Science Quarterly in 2016, sets aside the explanation by complacency or technical inferiority. The central cause is a shared and, above all, asymmetric fear. The executives feared external competitors and shareholders. The middle managers feared their own superiors and their peers.
That asymmetry produces a documented loop. The executives’ external fear leads them to press the middle managers without revealing the real severity of the threat, and to read the escalations back in a biased way. The middle managers’ internal fear lowers their propensity to pass on negative information. The top then develops an over-optimistic reading of its own technological capabilities and neglects the long investments.
The filtering is neither a choice nor a moral failing, it is the mechanical product of an attention structure. Correcting this pillar therefore means treating the fear of the middle managers rather than demanding more transparency, because that demand, coming from the people they fear, worsens exactly what it claims to solve.
The five-step sequence to install this pillar
This pillar installs in a precise order, and the organisations that fail almost always start at the wrong step. They create a collection apparatus before resolving the processing, which produces reports nobody reads. The order matters more than the tools.
The five steps in order
Each step assumes the previous one. An organisation that skips the first gets empty channels, and an organisation that skips the third accumulates signals without ever producing a decision.
- Map the filters. Between the front line and your strategic decisions, identify each hierarchical level and assess honestly whether it has a structural interest in passing contradictory signals on or in softening them.
- Build channels that bypass the identified filters. Direct access to the front line for the leadership, escalation possible without going through the reporting line, and strategic reviews that include unfiltered operational testimony.
- Create the processing body before you open the collection. Who assesses a signal, against what criteria, with what authority to trigger a response, and within what deadline.
- Make an explicit decision mandatory on every processed signal. Monitor closely, act, or set aside with the reasons documented. The absence of a decision is itself a decision to ignore it.
- Widen the scanning beyond sector boundaries, once the first four steps are in place. The most significant disruptions rarely come from inside the sector.
The most common sequencing mistake
Most organisations start at the fifth step, buying scanning tools and subscriptions to business intelligence services. It is the most visible step, the easiest to budget and the most useless while the previous four are missing.
Collection without processing produces a sense of vigilance without producing its effects. That feeling is even counterproductive, because it creates the illusion that the organisation is watching its environment when it is merely documenting it.
A processing body only turns a signal into a decision quickly when the authority, the amounts and the thresholds were settled in advance, which is the subject of structural responsiveness.
The order also explains why this pillar sits on top of two others. It needs a culture of internal dissent so the signal is spoken at all, and it needs long term governance so a signal that threatens this quarter is not automatically the one that gets set aside. A scanning apparatus built without those two produces full channels and empty decisions.
The signals you have an interest in not hearing
One diagnostic question is more uncomfortable than the others, and it points to the most expensive blind spots. What are the signals your organisation has a structural interest in not hearing? The answer is rarely comfortable and always useful.
Identify your structural blind spots
Every organisation has them. They are the signals that would question its dominant business model, its historical investments, its competitive positions or the skills on which the legitimacy of its leaders rests.
Kodak had an interest in not hearing that digital would make film obsolete. Boeing had an interest in not hearing what threatened its certification deadlines. In each case those signals existed without an organisational path, because the architecture implicitly rewarded their suppression.
The exercise is simple to run. Identify the three developments that would most fundamentally question your current model. For each, ask whether the teams that would perceive the early warning signs have an interest in escalating them or in playing them down.
Build sensors with no interest in the status quo
Where the interest is to minimise, an ordinary channel will never be enough. You need arrangements entrusted to people or teams with no stake in maintaining the current model.
Three forms work, and each addresses a different weakness. The first is external scanning entrusted to actors with no link to the incumbent model, so their reading is not shaped by what the current business protects.
The second is regular forums where speakers from outside the sector share their reading of developments. An organisation hears its own assumptions repeated inside its industry conferences, and the useful signal usually comes from an adjacent field that is two moves ahead on the same shift.
The third is an organised contradiction arrangement. Members of the organisation receive an explicit mandate to build the strongest possible argument in favour of disrupting the existing model, and the mandate matters because it converts a career risk into an assigned duty. The person arguing against the model is doing their job rather than sticking their neck out.
None of the three works as a one-off. A single external forum or a single red-team session produces a memorable afternoon and no lasting change to what the organisation hears. The arrangement has to recur on a fixed cadence, so that the uncomfortable reading becomes routine rather than an event.
Treat the cause before the symptom
Your signals travel badly and you are looking for why? Start with the pillar this one depends on entirely, the culture of internal dissent, without which no escalation channel ever produces anything except polite reports.
Which arrangement given your organisation
The right arrangement depends on the nature of the dominant filter in your organisation. Installing an anonymous channel in a structure where the problem is the processing rather than the escalation produces nothing, and it spends the credibility you need to install the right one afterwards.
Identify your dominant filter
A simple test names it. Take the last significant surprise your organisation suffered and trace its internal thread. The point where the signal was lost indicates the filter to treat first, and it is usually not the one leadership assumes.
| Where the signal was lost | The dominant filter | The priority arrangement |
|---|---|---|
| Nobody saw anything coming | No sensors in contact with reality | Designated sensors in each operational function |
| People knew and said nothing | Fear of personal consequences | Channel independent of the reporting line, with explicit protection |
| The signal rose then stopped | An intermediate level’s interest in not passing it on | Bypass the identified level and give direct access to the front line |
| The signal arrived but was not processed | No processing body and no decision authority | Monthly dedicated body with a mandatory explicit decision |
| The signal was processed then set aside | A structural blind spot tied to the business model | Scanning entrusted to actors with no interest in the status quo |
The special case of organisations that are doing well
An organisation that has suffered no recent surprise cannot run that test, and that is precisely the riskiest situation. The absence of surprise signals either an architecture that works or an environment still stable, and the two look identical from the inside.
In that case the test moves. Ask three operational managers at different levels to name a structural problem they perceive and that has never reached the executive team. Their answers are worth every diagnostic.
How I can help you build this pillar
This is the hardest pillar to diagnose, for a simple reason. An organisation that does not capture its weak signals generally does not know that it fails to. The blind spots are invisible from the inside, which is where an outside eye earns its place.
An audit of the escalation channels
The audit identifies, in a few weeks, the filters that distort signals before they reach decisions, the structural blind spots, and the existing mechanisms that are no longer used. It is often the most sober starting point, because it works on what exists rather than reinventing everything.
A workshop to design the scanning mechanisms
The workshop concretely builds the missing processes: alternative escalation channels, contradictory-signal forums and alert processing protocols. One working day that produces an operational apparatus ready to deploy rather than a report.
A keynote on organisational intelligence
The keynote is for organisations that want to put this question at the scale of a leadership team or an ecosystem. It draws on documented cases and produces a collective realisation about what the organisation already knows without yet hearing it.
The content sits on the page for the keynote on the invulnerable company. If you want to talk through what this would look like inside your organisation, book a conversation with me directly.
Conclusion
The organisations that come through crises without deep damage are not the ones that got lucky. They are the ones that built mechanisms to hear what they had a structural interest in not hearing.
Boeing knew. Kodak knew. Nokia knew, and the research measured why that knowledge never reached a decision. In all three cases, what was missing was not individual foresight but the architecture that turns foresight into collective action.
An organisational surprise is almost never a surprise. It is information that existed and did not find its path.
Capturing weak signals is not a refinement for ambitious organisations. Curiosity for weak signals is a survival mechanism in an environment where disruptions do not wait for you to be ready.
Frequently asked questions about weak signals
What is a weak signal in business?
Information already available in the organisation but under-interpreted, often carried by operational teams well before the threat becomes visible. As soon as a signal appears in a sector report it is no longer weak, and the advantage it carried has gone.
Is a scanning tool enough to capture weak signals?
No, and this is the most common sequencing mistake. Collection with no processing produces reports nobody uses, and a counterproductive sense of vigilance, because it masks the absence of any real apparatus for turning signals into decisions.
Why does bad news not travel upward?
The Nokia research shows the filtering comes from an asymmetric fear. Middle managers fear their superiors and peers, which lowers their propensity to pass on negative information, without any conscious decision to conceal being made.
How do we know if our organisation has blind spots?
Identify the three developments that would most question your current model, then ask whether the teams that would perceive the early signals have an interest in escalating or playing them down. Where the interest is to minimise, that is the blind spot.
Does this pillar work without a culture of dissent?
No. The two pillars are interdependent. Capture channels in an organisation that implicitly penalises contradiction produce polite escalations with no content, which gives the appearance of the apparatus without its effect.
