A supplier may need access to your systems to maintain equipment or process orders. If someone steals that supplier’s login, they may be able to enter through an approved connection. Keeping outsiders off the company network is therefore only part of protecting the business. The security perimeter, meaning the boundary between your network and the outside world, does not tell you everything about who can reach important information or stop work.
For the person approving security spending, the useful question is what the business would be unable to do if protection failed. Would orders stop leaving the warehouse? Could someone change payment instructions? Would staff lose access to customer records? A long list of security problems does not answer those questions. The team needs to connect each problem to the work, people and customers it could affect before you can judge which improvements deserve attention first.
Follow an order through the business
Choose an important activity and ask the people doing it to show how it works. For an order, that might mean following the customer’s request through approval, production and dispatch. Identify who can enter or change information, which systems staff use and where a supplier takes over. Include the less visible steps: someone transferring a file, approving an unusual request or using a supplier login. Each can matter as much as the main application.
Then ask what happens if a system stops, gives staff incorrect information or lets the wrong person see it. Could staff continue by hand, and for how long? Check what they need to restart as well. A login service checks who a user is and what they may access. If staff need that same service to sign in to their backup systems, losing it could stop both normal work and the planned recovery. This is an identity dependency: several parts of the business rely on the same way of proving who someone is. A small system can have a large consequence for that reason.
Understand what each improvement would prevent
Some security work needs to be maintained regardless of which new project wins funding. Know who is responsible for each system, limit access to the people who need it, keep software supported, look for signs of misuse and check that recovery works. Meet applicable legal and contractual requirements too. Choosing the order of additional work does not make a requirement optional.
Security teams call protective measures “controls”. Ask them to explain what each proposed measure would do in the failure you are considering. Requiring an extra check at login may make a stolen password less useful. Separating parts of the network may prevent an attacker on a supplier-connected system from reaching production. A recovery exercise answers a different question: can staff get orders moving again after an attack? These measures solve different problems and may be needed together.
Two measures do not necessarily provide twice the protection. Both might block the same route, leaving another open. They might also depend on the same login service and fail together if it is taken over. Ask engineers to show what the new measure adds to what is already in place. A system diagram and a relevant exercise can help establish that; a single security score cannot explain it.
Compare costs without hiding the uncertainty
The NCSC’s guidance on putting financial estimates on cyber risk recommends comparing the estimated reduction in the chance or cost of an incident with the cost of protection, while making uncertainty clear. For an interrupted service, consider earnings lost while work stops, the people needed to respond and the cost of getting back to normal. Compare options over the same period. Include ongoing maintenance and any disruption caused by installing the protection itself.
Estimates should show a range when nobody knows exactly how often an incident might happen or how much it would cost. Explain which assumption could change the spending decision. If the cheaper option depends on staff being able to work by hand, testing that process may help more than making the financial calculation more elaborate. Also examine severe incidents separately. An average loss estimate, which weighs possible losses by their likelihood, can hide an unlikely event that would leave the company unable to pay its bills or meet a safety obligation.
Make clear what waiting leaves at risk
If work is postponed, record what could still go wrong, which service would suffer and who has authority to approve that choice. Name any temporary protections the decision relies on, the person keeping them working and how they were checked. Set a date to reconsider. Review sooner if attackers start using the weakness, access changes or the system takes on different work.
A temporary measure reduces the danger without necessarily fixing its cause. Remediation means correcting the underlying problem and checking that the fix works. Keep that distinction visible: closing a ticket must not make a temporary workaround look like a completed repair. The manager responsible for the service and the team maintaining the workaround both need to know what remains at risk.
Ask for evidence that work can continue
At the next spending review, ask for more than the number of security tickets closed. Ask what an exercise showed about staff’s ability to keep the important work going or restart it. Did they have the access they needed? Could the alternative process handle the required orders? Bring overdue temporary arrangements back to the people who approved them. This gives you a basis for deciding what to fund next, rather than assuming that a shorter problem list means a safer business.