Digitalization promises faster decisions, simpler processes and better service. Yet the same systems can expose weak data, unclear ownership and hidden dependencies across the business. The central question is therefore not whether a company is becoming more digital. It is whether the company can still operate, control its records and serve customers when the technology changes or fails.
A digital programme begins with the work the business needs to perform, not with the software it plans to buy.
The system carries the rules of the business
Companies often start a digital programme by selecting a platform and then asking the organisation to fit around it. That may simplify procurement, but it reverses the more important sequence. Before choosing the software, management should be clear about the process, the data owner, the control and the business decision that the system is expected to improve.
An enterprise platform does more than process transactions. It decides who may approve a payment, how a customer is identified, when an order is complete and which figure is treated as the official result. If those rules are inconsistent before configuration begins, the new system will not resolve the disagreement. It will reproduce it at greater speed and scale.
The purpose also differs by sector. A manufacturer may need a dependable view of production and downtime. A retailer needs store, warehouse and online stock to agree. A utility needs meter, billing and network data to support the same customer account. A public body needs digital services that improve access without weakening financial control or accountability. The technology should serve these operating needs, rather than becoming the objective itself.
Management question: Before approving the platform, can management name the service, decision and control that will improve?
TSB shows that moving the data is not enough
In April 2018, TSB Bank moved its corporate and customer data to a new IT platform. According to the Financial Conduct Authority, the migration itself succeeded. The platform then suffered technical failures that disrupted branch, telephone, online and mobile banking. All branches and a significant proportion of the bank’s 5.2m customers were affected at the outset.
The regulators later fined TSB £48.65m for failures in operational-risk management, governance and oversight of outsourcing. The bank also paid £32.7m in customer redress. This case matters because the technical milestone was achieved: the data moved. The business outcome still failed because the wider service was not ready to absorb the change.
A migration plan may show that records reconciled and interfaces passed their tests. Management must also know whether customers can complete essential tasks, whether staff can work when a channel fails and whether suppliers can support recovery at the speed the business requires.
Management question: If the new platform fails on its first busy day, which services must remain available and who can activate the fallback?
Birmingham shows the cost of losing a reliable record
The troubled Oracle implementation at Birmingham City Council offers a different warning. Reporting on the programme described extensive manual correction, unpaid suppliers, uncollected debts and difficulty producing dependable financial information after the system was introduced. The council has continued to spend heavily on remediation and reimplementation.
This is more than an accounting inconvenience. When the finance system cannot provide a trusted view of cash, liabilities, debtors and suppliers, managers lose information needed to control the organisation. Staff then create spreadsheets and manual checks around the system. These workarounds may keep individual tasks moving, but they create new versions of the truth and make reconciliation harder.
The warning signs usually appear before go live: repeated data-cleansing delays, unresolved process exceptions, tests based on tidy samples, unclear ownership between the organisation and its supplier, or employees trained on screens but not on the changed process. A credible programme needs an agreed threshold for postponing go live. Without one, the date can become more important than readiness.
Management question: Could finance produce a trusted balance and pay critical suppliers on the first day after migration without rebuilding the answer in spreadsheets?
Digital resilience depends on calm preparation, clear priorities and recovery procedures that work outside the technology team.
Connected companies also inherit connected failures
Once a platform is live, a second risk becomes visible. Stores, payments, suppliers, employees and customers increasingly depend on shared networks, cloud services, identity systems and software updates. This improves efficiency, but it also allows one failure to interrupt several functions at once.
Cyber risk is therefore not limited to the theft of data. An incident can stop orders, prevent payments, disrupt warehouses and remove access to management information. The same operational damage can result from a defective update or an unavailable provider even when no attacker is involved. Resilience begins by identifying the services that matter most, how long they can be unavailable and which dependencies could stop them.
Management question: Does the board know which digital dependency could stop revenue, service or financial control within one hour?
Marks and Spencer made recovery a business wide task
The 2025 cyber incident at Marks & Spencer did not remain inside the IT department. Its 2026 annual report says the incident affected trading and product availability through the year, required frequent board attention and reset parts of the digital and technology transformation programme.
The financial statements recorded £131.3m of direct system-recovery, risk-management and specialist-advisory costs. Of this, £109.3m related to the immediate response and recovery. The external auditor treated the incident as a key audit matter because of its broad effect on operations and the control environment.
M&S restored systems progressively, rebuilt parts of its technology estate and used interim controls while normal processes were unavailable. That work illustrates why recovery cannot be left to the cyber team. Finance must protect the integrity of records, operations must keep goods moving, customer teams must communicate and leaders must decide which services return first.
Management question: If core systems were unavailable tomorrow, which manual controls would protect cash, stock, suppliers and customer data?
Digital value emerges when isolated tools become a secure, observable and coordinated operating system.
CrowdStrike shows how speed amplifies a small defect
On July 19th 2024, CrowdStrike released a content configuration update to Windows sensors. Its root-cause analysis found that one part of the system expected 21 input fields while another supplied 20. The resulting memory error caused affected Windows systems to crash.
The faulty file was identified and reversed within 78 minutes, but the update had already reached connected systems around the world. Some machines required direct attention before they could recover. A short deployment window therefore created a much longer operational problem for customers.
CrowdStrike responded with stronger checks, expanded testing, staged deployment rings and more customer control over when rapid-response content is released. The wider lesson is clear: when a change can reach thousands of critical devices quickly, validation, limited releases and rollback are business safeguards, not technical formalities.
Management question: Can a high-risk update be stopped after the first small deployment group, or only after every critical system has received it?
The same discipline should govern change and recovery
The four cases point to one story. Digitalization changes how the business operates, while resilience protects that operating model when technology does not behave as expected. Both require management to look beyond project schedules and security certificates. The company needs evidence that real data reconcile, complete customer and supplier journeys work, peak volumes have been tested and practical alternatives exist.
Third-party services deserve particular attention. A contract can transfer responsibility for delivering a service, but it cannot transfer the operational impact when the service stops. Apparently separate suppliers may also depend on the same cloud region, identity provider or network. Unless these links are mapped, a company can buy several services and still retain one point of failure.
Recovery exercises should involve operations, finance, communications, legal, customer service, technology and senior management. The aim is not to prevent every incident. It is to make sure that one incident does not become the whole company’s problem.
Management question: Has the company tested recovery as an operating business, or only tested whether its servers can be switched back on?
Where experienced digital advice adds value
An experienced adviser connects the business case, process design, data controls, system architecture, supplier obligations and operational testing. Difficult questions are raised while the answers are still affordable: before the platform is configured, before the go-live decision and before an incident forces the organisation to discover its dependencies under pressure.
The practical test is straightforward. A digital investment should make the company easier to operate and control, not merely more modern to describe. The strongest digital strategy is not the one with the most connected systems. It is the one that can continue serving customers, protect reliable records and recover in a controlled order when a connection fails.
Explore related expertise
Explore our data management and risk assessment and internal controls services. For a related perspective, read From Connected Assets to Better Industrial Decisions.