Information Decay#

“The single biggest problem in communication is the illusion that it has taken place.” — George Bernard Shaw

Every organization is, at its core, an information-processing machine. Raw data enters at the edges — from customers, from the market, from day-to-day operations. It flows upward through layers of management, getting summarized, interpreted, and filtered at each stop. By the time it reaches the person at the top, it’s been transformed. The question is: transformed into what?

In healthy organizations, information gets compressed but not distorted. The essence survives the trip. In dying organizations, information decays — each layer between the front line and the decision-maker strips away nuance, softens the bad stuff, and pumps up the good stuff until the picture at the top looks nothing like reality.

This is the closing chapter on strategic decision failures. The previous three chapters looked at direction errors, autocratic risk, and the cost of hesitation. This one addresses the foundation beneath all of them: the quality of the information decisions are built on.

Case 1: Ironclad Security — Five Layers Between the CEO and the Customer#

The Rise#

Ironclad Security was a commercial security services company founded in 2004 by a former law enforcement officer. The company provided armed and unarmed guards, surveillance monitoring, and access control for corporate campuses, shopping centers, and residential communities.

By 2012, Ironclad had grown to 2,000 employees and $85 million in revenue across eight states. It held long-term contracts with several Fortune 500 companies. The founder-CEO was proud of the organization he’d built — a tight hierarchy modeled on the military chain of command he knew from his law enforcement days.

The Fall#

Ironclad’s org chart had five management layers between the guards on the ground and the CEO: shift supervisors, site managers, regional directors, VPs of operations, and the COO. Every layer had reporting duties. Every layer produced weekly summaries for the one above.

The problem wasn’t the structure itself. It was what happened to information as it climbed.

A security guard at a corporate campus notices that employees keep propping open a fire exit to smoke — a real security hole. He flags it to his shift supervisor. The shift supervisor logs it in his weekly report as “minor compliance issue at Site 14.” The site manager, pulling together reports from twelve shift supervisors, writes: “compliance generally good across all sites with minor exceptions.” The regional director, reading summaries from eight site managers, tells the VP: “operations running smoothly in the Northeast region.” The VP reports to the COO: “all regions performing within parameters.” The COO tells the CEO: “no significant operational issues.”

This wasn’t hypothetical. This was the actual chain of events that cost Ironclad its largest contract in 2014. A security breach at a Fortune 500 client — directly tied to the kind of vulnerability guards had been flagging for months — led to a theft of proprietary equipment worth $2.3 million. The client killed the contract and filed a negligence lawsuit.

When the CEO dug in, he found that front-line guards had documented the problem over and over. The information had entered the system. It just didn’t survive the climb.

Losing the contract and facing the lawsuit set off a chain reaction. Two other major clients ran their own audits and found the same patterns. Both walked. Revenue dropped 35% in eighteen months. The company was sold to a competitor in 2016 at a distressed price.

The Lesson#

Ironclad’s CEO didn’t ignore information. He never got it. The five-layer hierarchy he’d built — staffed with competent, well-meaning managers at every level — worked as an information filter that systematically stripped out exactly what mattered most: the anomalies, the edge cases, the uncomfortable truths.

Information decay isn’t caused by dishonesty. It’s caused by structure. Each layer of summarization is a layer of loss. The more layers between the decision-maker and reality, the less reality gets through.

Case 2: Broadfield Retail — The Dashboard That Replaced the Store Visit#

The Rise#

Broadfield Retail ran forty-five home improvement stores across the Midwest. Founded in 1995 by a husband-and-wife team, the company had grown steadily for two decades on the strength of knowledgeable staff, fair prices, and a deep feel for the local markets they served.

The founders spent their first fifteen years visiting every store at least once a quarter. They walked the aisles, talked to employees, watched customers, and adjusted strategy based on what they saw firsthand. That direct connection to the front line was the company’s real competitive edge.

In 2013, the founders brought in a professional CEO to manage the next growth phase. He came from a big national retailer and carried a data-driven playbook. Within a year, he’d rolled out a full business intelligence platform — dashboards tracking sales per square foot, inventory turnover, labor costs, customer traffic, and dozens of other metrics in real time.

The Fall#

The dashboards were technically impressive. They served up more data, faster, than the founders ever had. But they replaced something the data couldn’t replicate: the qualitative, contextual understanding that comes from standing in a store and looking around.

The new CEO managed by dashboard. He stopped visiting stores. His regional managers followed suit, cutting store visits from weekly to monthly to quarterly. Decisions about layouts, product mix, and staffing were made from behind a screen.

The numbers missed what eyes would have caught. A store in a college town was bleeding customers — not because of pricing or selection, but because a neighboring construction project had turned the parking lot into a nightmare. A store near a military base was stacked with high-end tools that nobody was buying, because the algorithm didn’t know about the upcoming troop rotation that would pull thousands of potential customers out of the area. A store in a rural community was short-staffed on Saturday mornings — the peak window — because the labor model optimized for average traffic, not actual peaks.

Every one of these problems was obvious to anyone who walked through the door. None of them showed up on the dashboard.

Over three years, same-store sales dropped 12%. The CEO’s response was to tweak the dashboard metrics and hire a data science team to build predictive models. The models were sophisticated. They were also wrong, because the underlying data had already been stripped of the context needed to make sense of it.

The founders came back to active management in 2018, but the damage ran deep. Seven stores were in the red. The company shuttered twelve locations and was eventually bought by a regional competitor in 2021.

The Lesson#

Broadfield didn’t fail because of bad data. It failed because of substitution — swapping direct observation for quantitative proxies that couldn’t capture what actually mattered. Dashboards are powerful, but they measure what’s measurable, not what’s important. The gap between those two is exactly where information decay lives.

Data isn’t understanding. A dashboard can tell you what’s happening. Only walking the floor can tell you why.

Case 3: Quantum Pharma Services — The Middle Manager Who Curated the News#

The Rise#

Quantum Pharma Services was a contract research organization founded in 2006. It managed clinical trials for pharmaceutical companies — recruiting patients, collecting data, ensuring regulatory compliance, and delivering results. By 2014, Quantum had 300 employees across four countries and was generating $55 million in revenue.

The founder-CEO was a scientist at heart — detail-oriented, exacting, and committed to making decisions based on evidence. She’d built the company’s reputation on data integrity and regulatory rigor.

The Fall#

In 2013, the CEO hired a Chief Operating Officer to handle the growing operational complexity. The COO was an experienced healthcare executive — polished, well-spoken, and politically sharp. He quickly became the CEO’s main window into day-to-day operations.

The COO had a particular gift: presenting information in the best possible light. When a clinical trial fell behind schedule, he framed it as “timeline optimization.” When a client was unhappy with data quality, he called it “an opportunity to deepen the relationship.” When a regulatory audit turned up deficiencies, he reported them as “minor documentation items, already handled.”

Nothing he said was technically untrue. All of it was misleading.

The CEO, who had always prided herself on evidence-based decisions, was now making those decisions based on evidence that had been curated by someone whose core skill was managing perception, not managing operations.

The reckoning came in 2016 when the FDA ran a comprehensive audit on one of Quantum’s biggest ongoing trials. The audit uncovered pervasive data quality problems — not fraud, but widespread sloppiness in data collection, documentation, and quality control. Project managers and QA staff had been flagging these issues internally. The COO had filed them under “operational improvement opportunities” and papered over them in weekly reports with phrases like “continuous improvement initiatives underway.”

The FDA put the trial on clinical hold. The pharma client pulled out and demanded a $12 million refund. Two other clients froze their projects pending their own reviews.

The CEO fired the COO, but the damage was done. Quantum’s reputation — its most valuable asset in an industry built on trust — was gutted. Revenue fell 40% over the next year, and the company was absorbed into a larger CRO at a fraction of its former value.

The Lesson#

The CEO didn’t hire a liar. She hired a skilled communicator who instinctively softened bad news and amplified good news. This behavior is so common in organizations that most people don’t even recognize it as information distortion. But when the person curating the news is the primary channel between the CEO and reality, the distortion becomes systemic.

The most dangerous form of information decay doesn’t come from too many organizational layers. It comes from a single trusted gatekeeper who tells you what you want to hear instead of what you need to know.

The Diagnostic Pattern#

This chapter — and the four-chapter arc on strategic decision failures — closes with a fundamental point: every decision failure is, at its root, an information failure.

• Fatal directions (Chapter 15) come from acting on assumptions that haven’t been checked against current reality.
• Autocratic decisions (Chapter 16) come from silencing the voices that carry inconvenient truths.
• Missed windows (Chapter 17) come from information processes that move slower than the markets they’re trying to track.
• Information decay (this chapter) is the foundation under all three — the slow corruption of the data that decisions depend on.

The three cases show three distinct decay mechanisms:

1. Structural decay (Ironclad): Too many layers between the front line and the decision-maker, each one sanding off nuance.
2. Substitutional decay (Broadfield): Replacing direct observation with quantitative proxies that can’t capture context.
3. Curatorial decay (Quantum): A trusted gatekeeper who filters information based on what the decision-maker wants to hear.

The diagnostic questions for any organization:

• “How many layers does information travel through before it reaches me?” Each layer is a potential point of distortion.
• “When did I last observe our operations directly — not through reports, not through dashboards, but with my own eyes?” If the answer is months ago, substitutional decay is likely already at work.
• “Who controls what information I see, and what are their incentives?” If the person curating your information is rewarded for good news and punished for bad news, you’re not getting information. You’re getting a performance.

The fix for information decay isn’t more data. It’s shorter paths between the decision-maker and reality — fewer layers, more direct observation, and a culture where delivering bad news is treated as a contribution, not a career risk.

Ultimately, the quality of your decisions can never exceed the quality of your information. And the quality of your information depends not on how much you collect, but on how little gets lost between collection and comprehension.