The Ambulance That Cried Wolf.
SMS built thirty years of urgency into a single notification sound. Brands are spending it on flash sales. The fraud alert arrives into the silence that follows.
On the afternoon of 14 August 2003, the largest blackout in North American history began with a software bug in an alarm system in Ohio.
The Firstenergy control room had been receiving alarms all morning. Not the blackout alarm — routine alarms, maintenance flags, status notifications that the monitoring software generated throughout any normal operating shift. The operators had seen so many of them that they had developed, without deciding to, a practical working assumption: most alarms do not require immediate action.
When the software bug silenced the critical alarm at 14:14 — the one indicating that three high-voltage transmission lines had failed — the operators did not notice its absence. The alarm system had trained them, through months of routine signals that turned out not to matter, to treat its outputs as background rather than as events requiring immediate attention.
By 16:10 that afternoon, 55 million people across eight American states and the Canadian province of Ontario had lost power.
The alarm that mattered had been communicating through a system that had cried wolf so many times its operators had stopped listening. Not dramatically. Not consciously. They had simply, and entirely rationally, recalibrated their internal threshold for what counted as a signal worth acting on.
The Research That Named This Mechanism
The human brain does not treat each new signal as independent. It uses the historical ratio of true signals to false alarms from any given source to set its detection threshold — the level of evidence required before classifying the next signal as worth acting on.
This finding was formalised in 1954 by American psychologists Wilson Tanner and John Swets, who introduced signal detection theory into experimental psychology from its origins in radar engineering. Their framework — extended by David Green and John Swets in their 1966 book — identified four possible outcomes in any detection task: a hit, a miss, a false alarm, and a correct rejection.
The critical insight was this: a source that produces many false alarms — signals that turned out not to require urgent action — causes the brain to raise its detection threshold. More evidence of genuine urgency is required before the brain classifies the next signal as worth acting on.
The brain has not become less capable of detecting the signal. It has become more sceptical of that particular source's signals, based entirely rationally on everything that source has communicated before.
This is what your brand's SMS sender ID is doing to your customers' brains right now.
The Urgency Budget That Nobody Manages
SMS is not a neutral delivery channel. It carries thirty years of accumulated meaning.
The channel's early use cases were almost uniformly urgent: OTPs expiring in sixty seconds, bank alerts requiring immediate attention, fraud notifications demanding a response before the window closed. An entire generation of smartphone users developed an internal detection threshold for SMS that was calibrated to a very high ratio of true signals to false alarms. The channel trained the brain to treat its signals as worth responding to immediately.
Then brands arrived with a perfectly measurable observation: SMS has 98% open rates within three minutes of receipt. The channel selection appeared obvious.
What the open rate did not measure was the detection threshold it was quietly raising with each promotional send.
Every SMS that carried manufactured urgency — the 24-hour flash sale that runs every other week, the "last chance" offer that always has another chance, the festive discount that arrives every festival — registered as a false alarm against that sender's detection history. The brain that opened the message looking for something requiring action found something requiring nothing. It noted the result. It adjusted the threshold upward, fractionally, for the next signal from this sender.
Do this often enough and the threshold reaches a level where a genuine fraud alert — sent from the same sender ID, in the same format, at a similar time of day — has to clear a bar that the promotional messages built before it.
The ambulance that drove through the neighbourhood with its sirens on to announce a minor road closure every day for six months has a specific problem the day the actual emergency happens.
The Four Symptoms of a Depleted Signal Budget
"Our SMS Click Rates Have Halved in Eighteen Months and We Can't Explain It"
The first symptom of signal budget depletion. The channel is delivering. The open rate is still reasonable. The click rate — which requires the customer to actually read the message and make a decision about it — has declined sharply.
This is not a content problem or a targeting problem. It is the detection threshold moving. The customer who received forty-seven promotional SMS messages in the past year has a different internal calibration for this sender's messages than the one who received seven. The threshold for what counts as worth acting on has risen with every false alarm accumulated.
The operators in Ohio were still watching the screens. They had simply recalibrated what movement on the screen was worth responding to.
"Our OTP Delivery Rate Is Fine But Customers Are Calling to Say They Never Got Through"
The second symptom — and the one that should alarm a security or compliance team more than a marketing one.
OTP delivery rate is a technical metric. It measures whether the message was successfully transmitted to the network. It does not measure whether the customer treated the message as urgent the moment it arrived — or whether, having been trained by months of low-priority signals from this sender, they attended to it a minute or two later than the one-time window required.
The customer who calls to say the OTP did not work often received it. The detection threshold simply required them to consciously override a trained response before acting on it — and sometimes that override comes a few seconds too late.
The critical alarm was received by the monitoring system. The problem was what happened between receipt and response.
"Fraud Alert Open Rates Are Dropping While Fraud Incidents Are Increasing"
The third symptom is where signal budget depletion produces measurable harm to real people.
A bank using its SMS infrastructure for both fraud alerts and promotional offers is sending fraud alerts through a channel whose detection threshold — for that sender — has been raised by months of promotional false alarms. The customer who received the fraud alert at 2:17pm on a Thursday received promotional messages from this sender on Monday, Tuesday, and Friday of the previous week. Their brain has a threshold for this sender's signals. The fraud alert has to clear it.
Often it does not clear it quickly enough. Not because the customer is careless. Because the channel was trained, by the brand itself, to produce this specific response.
"We Switched to a New Sender ID and Saw Immediate Improvement — Then It Faded After Three Months"
The fourth symptom reveals the mechanism most clearly.
A new sender ID starts with a false alarm history of zero. The detection threshold for it is uncalibrated — the brain has no prior evidence about whether this source's signals are worth acting on, so it defaults to treating them as potentially important. Click rates recover. OTP response improves. The team concludes the new sender ID has solved the problem.
Three months later the decline reappears.
Because the promotional messaging behaviour did not change with the sender ID. The false alarms began accumulating from the first promotional send on the new ID. Signal detection theory does not care which number you use — it cares about what you have historically communicated through it.
A new sender ID is not a solution. A different ratio between urgent and non-urgent messaging is.
Three Decisions That Rebuild the Budget
Signal detection theory does not prescribe silence. It prescribes a different relationship between what a channel signals and what it delivers.
Decision 1 — Separate the urgency channel from the promotional channel
Transactional SMS and promotional SMS should never share a sender ID. A dedicated transactional sender ID, used exclusively for messages requiring immediate action, maintains a false alarm rate close to zero. The detection threshold stays calibrated to genuine urgency. The promotional channel operates separately, with its own sender ID, managed with its own frequency considerations.
This single structural decision preserves the signal budget of the transactional channel regardless of what happens to the promotional one.
Decision 2 — Set a promotional frequency limit as a budget constraint
The urgency budget is finite. It depletes faster than it accumulates. As a practical translation of the signal detection principle — not a research-derived rule, but a working operational guide — the promotional sends on any sender ID should be outnumbered by transactional sends on the same ID over any rolling 30-day period.
The specific ratio matters less than the discipline of maintaining one. The brands that treat SMS frequency as a budget rather than a volume opportunity are the ones whose channels still work when they need them most.
Decision 3 — Audit the false alarm rate before the next campaign
Pull the last twelve months of SMS sends from your primary transactional sender ID. Classify each send as transactional (required immediate action) or promotional (brand-initiated content).
If more than 30% of sends were promotional, the detection threshold for that sender ID has been materially raised. The next fraud alert is going out into a customer base whose brains have been trained to treat this source as lower priority than it needs to be.
That number is the diagnosis. The three decisions above are the treatment.
The Alarm That Got Through
The restoration of power after the 2003 blackout took between four hours and four days depending on location. The investigation that followed produced something more lasting than a repair.
The North American Electric Reliability Corporation established mandatory standards for control room alarm management that are still in force today — standards that specifically limit alarm flood rates, require suppression of low-priority alarms during high-activity periods, and mandate that critical alarms be visually and audibly distinguishable from routine ones.
The insight was not that fewer alarms should be sent. It was that alarms of different urgency levels needed different signal characteristics — managed deliberately, with a clear understanding of what each category of signal was worth — so that the brain of the person receiving them could maintain an accurate detection threshold for each type.
The power grid now manages its signal budget the way every SMS programme should.
Not fewer messages. The right messages in the right channels at the right frequency — so that when the signal that matters arrives, the person receiving it still knows exactly what to do with it.
Your customer's brain is the control room.
Your sender ID is the alarm system.
Manage it like the thing people's safety actually depends on — because for some of your customers, right now, it does.
If your brain is already triaging this page for a 5-second window, skip the reading—the complete narrative is perfectly laid out in the infographic below.
Published by Hetvabhas — independent analysis of brand communication
