1. Why dependencies determine real recovery times
A process does not stop only when the system it uses directly fails. It also stops when a service that system depends on fails: power, connectivity, the identity service, email, a payment provider or a courier. The real restart time is determined by the slowest dependency to reactivate, not by the most visible one. Anyone looking only at the main system tends to underestimate the problem, because the real failure lies upstream.
For this reason the mapping should be done before choosing recovery strategies. Without a map, the plan protects one link of the chain well and leaves exposed precisely the link that fails first. The map is a living document, and its first result is almost always a surprise for those who build it, because it reveals dependencies that nobody had ever mentioned in a meeting.
2. Prerequisites
Three things are needed to build the map. The first is the list of critical processes, already ranked by priority, derived from the impact analysis. The second is the inventory of systems and services used, including those not managed by the IT department, such as tools bought directly by business functions. The third is the time of those who really know the processes: without interviews with managers, the map remains a diagram of systems rather than of the activities they support.
Also prepare to gather contracts and documentation from external suppliers. Information on service conditions and responsibilities is often already written down somewhere, but nobody remembers it or has ever compared it with the actual processes. Re-reading the contracts with the list of processes in hand is one of the most useful pieces of work, because it relates two documents that usually live separately.
3. How to map them: the steps
Work process by process, starting from the most critical. For each one, follow the same sequence, so that the maps are comparable with each other and easy to update. A complex format is not needed: a shared table, with one row for each link, is sufficient for most medium-sized companies. The result should then be reviewed by whoever carries out the process every day, because they are the ones who recognise the dependencies that do not appear in any document:
- Write the process and the result it must produce when it works
- List the systems and the people who carry it out
- For each system, note which services it depends on, such as power, network, identity and integrations
- Identify the external suppliers involved and the service they provide
- State for each dependency who manages it and which recovery objective they declare
- Flag the points where a single person or a single access is required
- Compare the map with the managers and correct any differences
4. Common mistakes
The first mistake is mapping only IT. Many critical dependencies lie outside the department: a bank account linked to a single person, a supplier that holds the order forms, a maintenance contract that does not provide for out-of-hours work. The map must start from the process, not from the technical role of whoever compiles it, because it is the process that stops, not the server.
The second mistake is treating dependencies as static. A new supplier, a migration or a person changing role alters the chain, and the map ages quickly. The third is stopping at the first direct dependency: the email service may depend on connectivity, which in turn depends on a line provider whose contract nobody remembers any more. Following the chain to the end requires some patience, but this is where the most serious risks are found.
5. How to check the map is useful
A map is useful when it allows you to make a decision. A practical check is a paper simulation of a failure: choose an external service and ask the group which processes would stop, after how long, and what could be done in the meantime. If the answers differ from person to person, the map is incomplete and must be completed before proceeding with planning.
Also verify individual links with whoever manages the service. Ask the supplier whether its dependencies on other services are documented and whether it will inform you in case of an outage. The information gathered this way must be carried into the continuity plan, not left in the memory of the person who obtained it, because that person may change role. The document must outlive whoever manages it.
6. What to do after the map
The map feeds three decisions. In the continuity plan, it becomes the list of points to protect and of risks to accept consciously. In supplier contracts, it becomes the list of clauses to request: declared timeframes, notification of outages, liability in the event of a chain failure. In technical projects, it becomes the criterion for deciding where redundancy is needed, avoiding spending where the chain is already robust.
Not all dependencies can or should be eliminated, and often it is neither possible nor cost-effective. Some are accepted, provided they are known, documented and periodically reassessed, with the date of the assessment. The difference between an accepted risk and an ignored risk is precisely the map, and management should be able to see which of the two is governing. Presenting it to management with this distinction makes it easier to obtain the decisions needed.
7. How to start
This week, choose the most critical process and map it down to external services, using only the information you already have: contracts, the list of systems and conversations with the manager. Clearly flag any dependencies that nobody can answer for, because they are the first to clarify. Even an incomplete map is a better starting point than no map at all, provided the gaps are declared.
Then arrange a meeting with the connectivity and power suppliers, and ask in writing which outages they foresee, who the contact is in case of a problem and what timeframes they declare. These answers are the first solid data point in the map and replace assumptions. Record the answers in the document together with the date you received them. If a supplier does not respond in writing, note that too, because it is already a piece of data about their availability.
Frequently Asked Questions
What is the difference between a critical dependency and an important one?
A dependency is critical when its interruption stops or damages a process that falls within the approved priorities. An important dependency may improve the work but does not block anything essential. The distinction depends on the process, not the system: the same service can be critical for one activity and secondary for another.
Who should update the map?
A manager should own it, usually whoever handles continuity or risk management. Updates should be gathered from whoever introduces the changes: IT for systems, purchasing for suppliers, operational functions for processes. The owner checks that changes are recorded and that the map remains consistent.
Do I need a dedicated tool?
Not necessarily. A shared table with fields for process, system, dependency, supplier, objective and owner is sufficient for many medium-sized companies. A dedicated tool can help when dependencies are numerous or change often, but the work of gathering and verifying with people is still necessary in every case.
How do you manage dependencies on very large suppliers?
You often cannot negotiate specific terms with large suppliers. In this case it is best to understand which guarantees they declare in their standard contract, what their communications are in the event of an incident and which alternatives you can prepare locally. The map also serves to document this dependency as an accepted risk, with the date of the assessment.
Should dependencies also be considered from a GDPR perspective?
The GDPR applies when a supplier processes personal data on your behalf. In that case a processing agreement is needed that governs security, sub-processors and breach management, and the map helps you know which suppliers process data. For specific obligations, check with your privacy adviser.