Every growing team hits the same wall: one person becomes the answer to every question.
They're the only one who knows how the billing system works. The only one who can configure the API. The only one customers ask for by name.
This isn't a compliment — it's a crisis waiting to happen. And the data shows just how widespread the problem is.
The Numbers on Knowledge Silos
The 2024 StackOverflow Developer Survey found that 45% of developers report that knowledge silos negatively impact their productivity three or more times per week. Not occasionally — three or more times per week. That's a constant drain on productive work.
Source: Enginuity Newsletter - Knowledge Silos accessibility.link.new-tab
Three independent studies converge on the same finding: 19-30% of working time is lost to knowledge inefficiency.
McKinsey (2012) found employees spend 1.8 hours per day (9.3 hours per week) searching for information — 20% of the workweek.
IDC found 2.5 hours per day (~30% of workday) lost to information search, with 60% of executives citing it as a barrier.
But the Panopto study (2018, n=1,001 US employees) breaks it down into its component parts:
5 hours per week waiting for the one colleague who knows
8 hours per week searching independently through trial and error
6 hours per week duplicating work that already exists somewhere
Total: up to 19 hours per week — 2.4 full workdays out of 5.
Sources: Glean accessibility.link.new-tab, Haiilo accessibility.link.new-tab, Helpjuice accessibility.link.new-tab
The 5-Hour Queue: Waiting for the Expert
That 5 hours per week waiting for "the one colleague who knows" is the expert bottleneck in action. It's not that the knowledge doesn't exist — it exists in one person's head.
In software, this is called the "bus factor" — the number of team members who would need to leave before a project is crippled. A bus factor of 1 means one resignation, one illness, or one vacation can halt operations.
Sources: Swimm accessibility.link.new-tab, Wikipedia accessibility.link.new-tab
The Financial Cost
The Panopto research found that the average large US business loses $47 million per year in productivity due to inefficient knowledge sharing. For organizations with 30,000 employees, that number climbs to $72 million annually.
And when the expert leaves? The Panopto finding that hit hardest:
42% of institutional knowledge is NOT shared with coworkers — meaning when that employee leaves, their coworkers cannot do 42% of their job.
Sources: Iterators accessibility.link.new-tab, Learn to Win accessibility.link.new-tab
Why Documentation Fails
The usual response is "let's document everything." But documentation assumes the reader knows what question to ask.
That 42% of unshared knowledge? It's unshared not because people are hoarding it — it's unshared because you can't search for knowledge you don't know exists.
The expert knows the shortcuts, the edge cases, the "this looks wrong but it's actually right" situations. A wiki entry can't capture the 15 years of context in their head.
Cognitive Load Theory (Sweller, 1988) shows that structured learning paths with worked examples consistently outperform reference documentation for skill acquisition. The difference is transformation vs. reference.
What Teams Actually Need
Organizations implementing comprehensive silo-busting strategies report up to 55% productivity improvements, with McKinsey finding that knowledge sharing improves productivity by 35%.
The pattern that works: converting expert knowledge into structured, accessible learning — not just documentation, but education that transforms novices into capable practitioners.
The person who knows the answer shouldn't have to answer it 50 times. But they also shouldn't have to write a 10,000-word wiki that nobody reads. They need to teach — and the format matters.
Your Nervous System and the Expert Bottleneck
If you're the expert everyone depends on, this isn't just an organizational problem — it's a physiological one.
Savic et al. (2018) showed that chronic stress causes measurable changes to the prefrontal cortex — the part of your brain responsible for decision-making, working memory, and cognitive flexibility. The constant interruptions of being everyone's answer drain the same cognitive resources you need for your actual work.
And Pencavel's Stanford research (2014, peer-reviewed) found that output flatlines after 50 hours per week — working 70 hours produces the same output as working 56. If you're staying late to catch up on "real work" after a day of answering questions, you're likely in the flat part of that curve.
Sources: Savic et al. (2018), Cerebral Cortex accessibility.link.new-tab, Pencavel (2014), The Economic Journal accessibility.link.new-tab
The solution isn't working harder. It's building systems that let your knowledge scale without requiring your constant presence.