Duskbridge / Group operations / Key-person risk / Operator notes
What to write down so your company does not live in one person's head
Write down five things, one page each, so your company's knowledge does not live in one person's head: where things live, recurring jobs, decision reasons, owners and backups, and critical dates and contacts.

Write down five things: where your important things live, how the recurring jobs get done, why the big decisions were made, who owns each job and who is the backup, and which dates and outside contacts matter. Keep each one to one page, and give each one an owner, a backup, and a review date. The five-item list is our own opinion, built from sources that point the same way.
Chances are your company has someone like this. They know which account renews in March, why the old vendor was dropped, and how to fix the thing that breaks every quarter. It is often a compliment: they were good, so the work piled up on them. Business advisers call the exposure key-person risk. It shows up on the day that person is sick, on vacation, or gone.
Why this is worth the effort
Researchers have a blunt name for it. In software, the “truck factor” is the number of people who would have to be hit by a truck, or quit, before a project is in serious trouble. A 2016 study by Guilherme Avelino and colleagues estimated it for 133 popular projects on GitHub, a site where programmers share code. In 87 of them, 65 percent, the number was two or fewer. In 45 of them, 34 percent, it was one. That is software, not a general business, so treat it as an example of the pattern and not a count of anyone’s company.
The Red Cross makes a similar point for businesses in its Ready Rating program. It says every decision maker, such as the person who can approve or reach emergency funds, should have at least one backup, and that planning should not stop at the executive level. People with specialized training or equipment skills need backups too. It also says more than one person should be able to recover systems such as email and reach stored data.
Writing things down is not a cure on its own, but there is some evidence that a structured handoff helps. In a before-and-after study of nine hospitals’ pediatric residency programs, published in the New England Journal of Medicine in 2014, a handoff program called I-PASS was followed by a 23 percent drop in medical errors, from 24.5 to 18.8 per 100 admissions, and a 30 percent drop in preventable injuries, from 4.7 to 3.3 per 100 admissions. The program combined a standard format with training and coaching, and the hospitals were compared with their own earlier results, so treat it as a signal and not as proof. A hospital is not a holding company, so this is an analogy. The lesson we take from it is small and practical: when knowledge has to move from one person to another, a clear, repeatable format helps.
The five things to write down
This list is our opinion, not a standard. The sources point the same way, but we chose these five.
| What to write down | What it looks like | Why it matters |
|---|---|---|
| 1. Where things live | A list of accounts, shared files, and who holds the admin rights. Never the passwords themselves | A second person can get in when the first is out |
| 2. How recurring jobs get done | Numbered steps for each job that repeats, with the decision points marked | A backup is less likely to miss a step |
| 3. Why big decisions were made | A one-page note: the situation, the choice, and what it led to | Newcomers can tell when a choice should be revisited |
| 4. Who owns what | Each job with an owner, a backup, and who can approve | Less guessing about who decides |
| 5. Dates and outside contacts | Renewals, deadlines, vendors, and who to call | Less chance a renewal lapses because one person forgot |
1. Where things live. Write down where each important thing is kept and who can open it. Do not write the password. Write which tool holds it and which two people can reach it. The Red Cross advice about having more than one person who can recover systems and reach stored data is the reason.
2. How recurring jobs get done. The Red Cross says to document essential processes and each person’s tasks in enough detail that a backup does not miss important tasks or decision points. In plain words: write the steps for the jobs that repeat, and mark the spots where someone has to make a judgment call. Our rule of thumb is to start with anything that happens more than twice a year.
3. Why big decisions were made. Software teams have a habit called an architecture decision record. Michael Nygard described it in 2011 for a simple reason: without the reasons behind past choices, newcomers can only accept them blindly or reverse them blindly, and both can do damage. His format has five parts: a title, a status, the context, the decision, and the consequences. He says the whole thing should be one or two pages, because short documents are more likely to be kept up to date. Martin Fowler’s site adds a useful rule: when a decision changes, do not rewrite the old record. Mark it superseded and link it to the new one. You do not need software to borrow this. A short note on why you picked a vendor, a price, or a policy is enough.
4. Who owns what. The Ready.gov business continuity template, from the United States government, asks for roles and responsibilities, lines of authority, who takes over if a manager is out, and who is allowed to make decisions in their place. For each job, write an owner, a backup, and who can approve.
5. Dates and outside contacts. The same template includes a team contact list and a table of vendors and contractors with a contact name and phone number for each. Add the dates that bite: renewals, filings, contract ends, and anything that stops working if it lapses.
Keep it small enough to survive
Nygard’s 2011 argument about short documents is the best reason to start small. His limit is one or two pages. Our version is stricter: one page per thing, so it can be written in about an hour and read in about five minutes. Those two times are our guess, not research.
Here are the habits we suggest. These are our ideas, not research findings.
- One page per thing. If it runs past one page, split it into two.
- An owner, a backup, and a review date on every page. Put the date at the top. We suggest 90 days.
- Keep it where two people can reach it. A note on one person’s laptop is the same problem in a different place.
- Practice with the backup. The Red Cross suggests using staff vacations as practice, letting the backup do the task, and then reviewing together what went wrong and recording the gaps. It also says to review backup plans at least once every 12 months, or sooner after the organization changes.
Worked example: a made-up company
Illustrative example, not Duskbridge data. A 12-person company has one person, Dana, who handles renewals, vendor accounts, and the monthly billing run. Dana is out for two weeks. The team opens a shared folder and finds nothing.
Here is what the rescue could look like, five pages in all.
| Page | What it says | Owner | Backup | Review |
|---|---|---|---|---|
| Where things live | The billing tool, the vendor portal, and the shared drive, and which two people hold admin rights | Dana | Marcus | Every 90 days |
| Monthly billing run | Eight numbered steps, with two marked "stop and check" | Dana | Marcus | Every 90 days |
| Why we switched vendors | The problem, the choice, and what changed | Dana | Priya | Every 90 days |
| Who owns what | Each job, its owner, its backup, and who approves | Priya | Dana | Every 90 days |
| Dates and contacts | Four renewals with dates, and six vendor contacts | Priya | Marcus | Every 90 days |
Nothing here is clever. The value is that Marcus can run the billing run while Dana is out, and Priya can see which renewal is due next. After Dana returns, the team uses a vacation week to let Marcus run the process from the page, and fixes the two steps that were unclear.
Three mistakes and the fix
Mistake 1: one giant manual. A manual that tries to cover everything is hard to finish and easy to leave unopened. The fix is one page per thing.
Mistake 2: writing it once. A page with no review date drifts away from how the work is actually done. The fix is the date on the page. When the date comes, the owner reads it, changes what is wrong, and resets the date. If nothing changed, that is worth knowing too.
Mistake 3: keeping the notes with the person. If the only copy lives on Dana’s laptop, nothing has changed. The fix is to store every page where at least two people can open it, and to name them on the page.
Where Duskbridge fits
Duskbridge is the parent company of Viceroy NM and Trunnion AI, the companies discussed here. As we described in our October 5, 2026 post, each company keeps its own brand, its own profit and loss, and its own leadership. That setup shapes how this idea works for a group, and the next paragraph is our opinion.
In a group, the natural rule is that each company writes down its own operating notes, because its own leaders know the work, and the parent writes down only what is shared. The same five questions apply at each level. The difference is who the owner and the backup are. This post is a method that any company can use. It is not a description of how any one company in our group keeps its records.
The review-date habit also appears in our November 8, 2026 post on how a small group decides what not to build, where each not-now entry carries one.
Run this week
- List the five or six jobs that would stall if one particular person were out for two weeks.
- Pick the one that would hurt most and write its steps on one page.
- Write down where things live for that job, and name two people who can reach them. Do not write any password.
- Put an owner, a backup, and a review date 90 days out on the page, then store it where both can open it.
- Ask the backup to run the job from the page the next time the owner is away, and fix anything unclear.
Talk to us
If this topic matters to your own company, or you want to compare notes on how a group of companies keeps its knowledge from living in one person’s head, contact Duskbridge.
This article is for general informational and educational purposes only and is not legal, financial, medical, or other professional advice. Duskbridge makes no guarantee that it is accurate, complete, or current. Use of this information is at your own risk.
Details current as of November 14, 2026. Next review February 12, 2027.
Sources
Sources accessed November 14, 2026.
- Guilherme Avelino, Leonardo Passos, Andre Hora, and Marco Tulio Valente, A Novel Approach for Estimating Truck Factors, arXiv 1604.06766 (2016); conference paper.
- American Red Cross Ready Rating, Back Up Strategies for Key Personnel.
- Amy J. Starmer and colleagues, Changes in Medical Errors after Implementation of a Handoff Program, New England Journal of Medicine, 371(19):1803-1812 (2014); PubMed abstract and Harvard Medical School summary.
- Michael Nygard, Documenting Architecture Decisions, Cognitect (November 15, 2011).
- Martin Fowler, Architecture Decision Record.
- Ready.gov, Business Continuity Plan.
- Duskbridge, What a parent company does for its companies, and what it leaves to them (October 5, 2026).
- Open Risk Manual, Key Person Risk.
- Duskbridge, How a small group decides what not to build (November 8, 2026).
This article may discuss strategy, possible capabilities, acquisitions, investments, or market development. Those statements are forward-looking, uncertain, and not commitments or guarantees; see Terms section 6.
Continue the conversation

