12 Knowledge Management Statistics for 2026

What the research says about the hours lost to finding things, and about the knowledge that walks out the door

Every company knows more than it can find. The answer to a customer’s question exists, in somebody’s sent folder or a thread from March, and the person who needs it will spend twenty minutes looking before writing it again from scratch. The statistics below size that gap: how long people spend searching, what it costs, and how much of what an organization knows exists in exactly one head. The Brain in this+that is our attempt at the other approach, a knowledge layer filled by the communication a team already produces rather than by somebody remembering to write things down.

Key Takeaways

• People whose jobs are mostly communication and coordination spend nearly 20% of the workweek looking for internal information, and 28% managing email, on the McKinsey Global Institute’s estimates.
• A large organization loses $47 million a year to inefficient knowledge sharing, on Panopto’s modeling for a 17,700-person company.
• 42% of institutional knowledge is unique to the individual who holds it, so it leaves when they do.
• Employees waste 5.3 hours a week waiting for information from colleagues or recreating knowledge that already exists.
• Median employee tenure is 3.9 years (BLS), the lowest since 2002, so knowledge loss is continuous rather than occasional.

How Long People Spend Looking for Information

1. Nearly 20% of the workweek goes to finding internal information

The McKinsey Global Institute report puts it at nearly 20 percent of the workweek spent “looking for internal information or tracking down colleagues who can help with specific tasks.” One day in five, locating things that already exist somewhere in the organization.

This is the origin of the 1.8 hours a day, or 9.3 hours a week figure that circulates widely. That version is the percentage converted into hours against a workweek longer than forty, rather than a number MGI published, which is worth knowing before quoting it.

2. And 28% of the workweek goes to email

From the same report: the average worker whose job is mostly communication and coordination spends an estimated 28 percent of the workweek managing email. Searching and email together take nearly half the week before any of the actual work begins.

3. IDC’s estimate is higher still, at 2.5 hours a day

IDC’s research puts knowledge-worker search time at 2.5 hours per day. The two estimates disagree by about 40%, which is worth stating plainly rather than picking the more dramatic one. Both describe the same phenomenon at the same order of magnitude.

4. Employees lose 5.3 hours a week waiting or recreating

Panopto’s research with YouGov found US knowledge workers wasting 5.3 hours every week either waiting for information from colleagues or recreating institutional knowledge that already existed. This measures something narrower than general search time: it is the cost specifically of knowledge that exists but is not reachable.

What It Costs

5. $47 million a year for a large organization

Panopto modeled the cost of inefficient knowledge sharing at $47 million annually for an organization of 17,700 employees. The figure is a model built on survey-reported hours rather than a measured loss, and it is best read per head.

6. About $2,700 per employee per year

Panopto’s $47 million applies to an average enterprise of 17,700 people, which works out to roughly $2,700 per employee per year. That division is our arithmetic rather than a number they publish, and it is the one that travels to organizations of other sizes: about $2.7 million at a thousand people.

7. $132.7 million a year at 50,000 employees

Panopto’s own upper figure is $132.7 million annually for a 50,000-employee business. The per-head cost does not improve with scale, which is the interesting part: bigger organizations do not get better at this, they get more expensive at it.

How Much Knowledge Has Only One Copy

8. 42% of institutional knowledge is unique to the individual

Panopto found that 42% of institutional knowledge is unique to the person who holds it. Not documented anywhere, not known by a colleague. Nearly half of what a company knows has a single point of failure attached.

9. Six in ten employees struggle to get information they need

60% of employees report difficulty getting the information they need to do their jobs well. This is the everyday texture of the numbers above: not a crisis, just constant friction.

10. 85% call knowledge sharing critical to productivity

85% of employees agree that preserving and sharing unique knowledge in the workplace is critical to increasing productivity. Near-unanimous agreement that it matters, alongside 42% of it still living in one head, is the gap this whole field exists to close. The survey base was 1,001 US adults at organizations of 200 or more employees, each with at least five years of work experience.

Why the Problem Never Settles

11. Median employee tenure is 3.9 years

The Bureau of Labor Statistics puts median tenure at 3.9 years in January 2024, down from 4.1 years in January 2022 and the lowest reading since 2002. Any knowledge base that depends on the people who built it is being rebuilt on a four-year cycle.

12. 3.2 million Americans quit a job in a single month

BLS recorded 3.2 million quits in June 2026, a rate of 2.0%. Each one takes a share of that 42% with them. Knowledge loss is not an event a company can prepare for; it is a rate it lives with.

The Pattern Underneath

The numbers describe two problems that look like one. The first is retrieval: information exists and cannot be found, which is what search time measures. The second is capture: information never got written down at all, which is what the 42% measures. Tools mostly address the first and assume the second has been handled.

It usually has not, and the reason is structural. Writing things down is work with a delayed payoff, done by people whose deadlines are immediate. So documentation happens when someone has time, which is to say rarely, and goes stale immediately after. Turnover then removes the person who knew what the document left out.

What changes the shape of it is capture that does not depend on anybody remembering. The answer to the customer’s question was already written, in a reply somebody sent in March. The decision behind the pricing change was already explained, in a thread. The knowledge is not missing, it is unindexed, sitting in the message stream where the work happened.

That is the bet behind the Brain: a knowledge layer that fills from the communication a team already produces. We have written about the reasoning in the missing middle of company knowledge and about why hand-maintained models decay in the Semantic Web didn’t fail because ontologies were wrong.

How We Sourced This

The Panopto figures come from its Workplace Knowledge and Productivity Report, conducted with YouGov among 1,001 US adults at organizations of 200 or more employees. The cost figures in that report are models built on self-reported hours rather than measured losses, and we have labeled them as such.

Tenure and quits data are from the Bureau of Labor Statistics. The McKinsey and IDC search-time figures are widely repeated across the knowledge management industry; we have linked the best available sources and reported both rather than choosing the larger. Where estimates disagree, we say so.