What key-person risk looks like in marketing
Every team knows who to ask about the awkward things. One person runs the monthly export from the booking or sales system into the email tool. One person understands the tag container (GTMTag management system.) well enough to change it safely. One person is the only administrator on the main ad account, and the payment method is their company card. One person knows that the brochure has to reach the printer a set number of days before the mailing slot, because they learned it the hard way. None of this is written down, because it has never needed to be.
The marketing risk register names "the only specialist in a channel leaves" as a people risk. In practice the exposure is wider and quieter than a departure. It shows up during a holiday, a period of illness or a busy week, when a task that has always happened simply does not, and nobody notices until a downstream effect appears: an email sent to an out-of-date list, a broken tag left in place, a campaign paused because the card expired. The process hub makes the same point about heroics: teams that depend on a few people who can rescue anything are fragile in proportion to how much they depend on them.
Several pages already handle pieces of this. The tutorial on reviewing access to marketing tools confirms an owner and a deputy administrator for each tool. The automation governance guide names a deputy for each workflow. The tutorial on documenting marketing integrations finds connections that run on someone's personal credentials. The SOP tutorial writes down one task so it survives staff changes. What none of them does is look across the whole team for everything that depends on one person. That is what this guide is for.
Building a coverage matrix
A coverage matrix is a table of the team's recurring tasks and the accounts and systems behind them, with the people across the top. In each cell, record whether that person can do the task unaided today, can do it with a procedureThe written way of doing one recurring task: who owns it, what starts it, the steps in order, checks and exceptions, so anyone qualified gets the same result. or some help, or cannot do it. It records capability, not responsibility.
That distinguishes it from two related tools. A RACI matrix assigns who is accountable for and who does each decision or deliverable; it can show one name against a task without saying whether anyone else could step in. The skills matrix in the guide to hiring and developing marketers rates people across skills to plan their development; a person can be rated competent in analytics and still never have run the specific export the business depends on. The coverage matrix asks the narrower question: if this person were away tomorrow, who could do this task?
Build the task list from what actually recurs, not from job descriptions. Walk through a typical month and quarter: reports produced, lists refreshed, feeds updated, budgets paced, invoices approved, campaigns launched, content scheduled, approvals given, accounts renewed. Add the seasonal tasks that come round once a year, because those are the ones nobody remembers how to do. Then add a row for each account or system that needs an administrator: ad accounts, social pages, the email platform, the CRMThe system that stores an organisation's contacts, companies, deals and interactions, and the processes built around it., the website's content system, the tag container, the analytics property and the asset libraryThe single home for approved marketing files, holding one current copy of each with its version, rights, expiry, alternative text and usage notes..
| Column | What it records |
|---|---|
| Task or account | A recurring task, or a system that needs an administrator |
| How often | When it runs, including seasonal or annual tasks |
| What depends on it | What stops or goes wrong if it does not happen |
| Person by person | Can do unaided, can do with the procedure or help, or cannot do |
| Procedure | Whether a written procedure exists, and when it was last followed |
| Access | Whose name the access, payment method or credentials sit under |
| Last run by someone else | When anyone other than the usual person last did it for real |
Fill it in with the team, not about them. People are usually candid about what only they can do once it is clear the exercise is about protecting the work, not judging anyone. Filled in this way, the coverage matrixA table of a team's recurring tasks and the accounts behind them, recording who can do each one unaided today, so work only one person can do becomes visible. describes the team as it stands today, so it changes whenever people, tools or tasks do.
Accounts and rights held in one name
The most dangerous single points of failure are often not skills at all but access. An ad account, a social page or a domain registration administered by one person can be locked the day that person leaves or is unreachable, and recovering it can be slow and uncertain. A payment method on one person's card stops every campaign it funds when the card changes. An integration connected with someone's personal credentials stops when their account is closed.
The company pages guide makes the point for LinkedIn: keep more than one super admin so the page is never locked when someone leaves. The same principle applies across the stack, and the access review tutorial is the procedure for it: confirm an owner and a deputy administrator for each tool, replace shared logins with personal accounts at the narrowest role the work needs, and move integrations to lasting accounts rather than personal ones. The integrations tutorial adds the question of whose credentials each connection uses. This guide does not restate those procedures; it uses the coverage matrix's access column to find where they have not yet been applied. Each periodic access reviewA periodic check of who can use each marketing tool: every account matched to a current person, supplier or integration, with owners, deputies and narrow roles. then confirms that every tool still has its owner and deputy after people change roles or leave.
Knowledge that is not written down
Some dependence is on judgement rather than access. The person who knows which stakeholder must see a price claim before it goes out, which supplier needs a call rather than an email, why a particular audience is excluded from a campaign or which numbers in the monthly report need a footnote holds knowledge that no tool records. Lose them and the work continues, but worse, with errors that nobody recognises as errors.
Find it by asking the people in the matrix, for each task they alone do well: what do you check that is not in any procedure, what have you learned to avoid, who do you call when it goes wrong? Record the answers. Some belong in a procedure for the task. Some belong in the approval matrix, the campaign brief template or the intake form, where they apply to everyone. Some are context that a new person needs once, which belongs in the team's onboarding pack for new marketers. Only a small part is genuine expertise that takes time to build, and that is where cross-training earns its cost.
Deciding which gaps matter
A coverage matrix in a small team will show many cells with a single name. Not all of them need fixing. Judge each gap by its consequence, not by how uncomfortable it looks: what would stop if this person were away, how soon would anyone notice, how hard would it be to recover, and does it touch money, customer data or what customers see?
A task that stops something customers rely on, such as an export that feeds every booking confirmation or departure email, matters more than a weekly internal report that could wait. A gap that fails silently, like a scheduled refresh nobody monitors, matters more than one that fails loudly. Access held in one name on an account that would be slow to recover matters more than access to a tool any administrator could reset. Seasonal tasks deserve extra weight, because the person who knows them may not be around when they next come round.
| Gap | Ask | Usually means |
|---|---|---|
| Only one person can run a recurring task | What stops, and how soon would anyone notice? | High priority if it feeds customer-facing work |
| One administrator on an account | How would the business recover access? | High priority; fix with a second admin |
| Payment method in one name | What runs on it? | High priority where live campaigns depend on it |
| Integration on personal credentials | What breaks when that account closes? | Fix through the integrations register |
| Judgement nobody else holds | Where does it belong once written down? | A procedure, the approval matrix or onboarding material |
| A seasonal or annual task | Who last ran it, and is it written down? | Write it down while it is fresh |
Record the gaps that matter in the risk registerThe single table of what could knock a marketing plan off course, each risk with a warning sign, an owner and a response prepared in advance., with an owner, so they are reviewed alongside the team's other risks rather than left in a separate spreadsheet.
Fixes: a second admin, a procedure, cross-training or a supplier
Match each gap to the fix that suits it, starting with the cheapest that works.
A second administrator fixes access held in one name. It is usually quick, and the access review tutorial covers it. Use the narrowest role that would let the deputy keep the account running, and record it in the access register.
A procedure fixes tasks that are repeatable but undocumented. The SOP tutorial writes one, including the checks, the exceptions and the new-starter test. A procedure turns a task one person can do into one that anyone qualified can follow, and it is usually the right first fix for recurring technical tasks such as exports and reports.
Cross-training fixes tasks that need judgement or skill a procedure cannot carry on its own. Pair the second person with the first on real runs, then reverse the roles. Use the skills matrix and development plans in the hiring guide to fit it into each person's development rather than treating it as extra work.
A supplier can provide cover where training a second person is not realistic, which is common in a very small team. A freelancer or agency on a standing arrangement can step in for specialist work. The comparison of in-house and agency marketing sets out the trade-offs of buying capability, and the guide to working as a marketing team of one covers buying help without losing control when the whole team is one person. A supplier still needs the procedure and the access to be useful in an emergency. Where the supplier's own job is to build or run the operating procedures, the guide to hiring and evaluating a marketing operations consultant sets out what the business keeps so that the risk does not simply move outside.
Whichever fix is chosen, record the new arrangement: who covers what, with which access, following which procedure. Assign one accountable owner for each fix, and keep the coverage matrix as the record of where cover now exists. Some gaps are best accepted rather than fixed; say so in the register, with the reason, so the decision is visible.
Testing cover before an absence forces it
Cover that has never been used is a hope, not a fact. A deputy administrator who has never signed in may find their invitation expired. A procedure nobody but its author has followed usually has a missing step. A colleague who was shown a task once may not remember it months later.
Test each fix with a real run by the second person, while the usual person is available but not involved. That is the same test the SOP tutorial applies to a new procedure, a new starter following it from the document alone. The deputy signs in and performs a real administrative task; the cross-trained colleague runs the export or the report for real; the supplier completes a small piece of live work. Record when each test happened and what it found, fix what failed, and rerun cover for the tasks that matter most from time to time, because procedures drift and access lapses.
Planned absences are the natural moment to test. A holiday announced in advance is a chance to have the second person run the task for real with help available on return. Treat the measures of this work as measures of the team's resilience, not as a KPIA metric tied directly to an objective. for any individual. When someone does announce they are leaving, the work moves from cover to handover, and the tutorial on handing over work when someone leaves takes over while they are still there. Once they have gone and the post stands empty, the tutorial on covering an empty role decides what stops, what is covered and what is bought in until the replacement arrives.
A coach holiday operator finds its gaps (illustrative)
Ulvarend Coaches finds its key-person gaps
Ulvarend Coaches, an invented coach holiday operator, had a marketing team of 4: a marketing manager, a digital executive, a content and brochure coordinator and a part-time designer. After the digital executive was off sick for a week and no departure reminder emails went out, the manager built a coverage matrix of 31 recurring tasks and 12 accounts.
Three gaps stood out. Only the coordinator knew the print schedule for the annual brochure, including the deadline for the printer's mailing slot. Only the digital executive could safely change the tag container. And only the digital executive ran the weekly export from the booking system into the email platform, which fed every pre-departure email; when it did not run, customers travelling the following week received nothing. The matrix also showed that the main ad account had one administrator and was paid on the manager's own company card.
The manager ranked the export first: it touched every customer about to travel and failed silently. The digital executive wrote a procedure, then the coordinator ran the export from it for 3 weeks while the executive watched, and found 2 missing steps that were added. A second administrator was added to the ad account and the payment moved to a company account. The brochure schedule was written down as a procedure with dates worked back from the mailing slot. The tag container stayed with one person, covered by a standing arrangement with the agency that built it, and was recorded in the risk register with that arrangement as its response.
Two months later the digital executive took a fortnight's holiday. The coordinator ran both exports, and every pre-departure email went out.
Common mistakes
- Mistaking a RACI or a skills matrix for coverage. One records who is accountable, the other who could grow into a skill. Neither shows who can run a task tomorrow.
- Listing jobs instead of tasks. Key-person risk hides in specific recurring tasks and specific accounts, not in role titles.
- Forgetting seasonal tasks. The task that runs once a year is the one nobody remembers how to do.
- Treating every single name as urgent. Rank gaps by consequence and by how quietly they would fail, and fix the worst first.
- Leaving access in one name. A second administrator is usually the quickest and most valuable fix of all.
- Cover that has never been tested. A deputy who has never signed in and a procedure nobody else has followed will both fail when needed. Test with a real run.
- Waiting for a resignation. By then the work is a handover under time pressure. Cover is cheaper while everyone is still there.
- Fixing it once. Teams, tools and tasks change. Update the coverage matrix when they do, and retest the cover that matters most.
