A proposed digital investment is usually examined line by line. Leaders ask about implementation, licences, internal capacity, disruption and overrun. The alternative may receive a few words: maintain the current position and revisit next year.

That is an asymmetric decision. One option is stress-tested; the other is treated as free.

Deferral can be the right choice. A weak proposal does not become sound simply because old systems have costs. The board still needs to compare two real courses of action, each with expenditure, risk, opportunity and consequences over time.

The source article called the hidden burden a “platform tax”. That remains a useful description. Maintenance invoices are only the visible part. The tax also appears as developer time spent patching, editorial work abandoned because every change needs support, duplicated manual processes and decisions constrained by what the current platform can accommodate.

“Do nothing” is rarely the actual option

Start by naming the alternative accurately. It might be:

  • operate the existing platform for another year;
  • renew a contract and accept a known limitation;
  • fund only security and continuity work;
  • defer one programme while improving the underlying process;
  • run a smaller piece of discovery before committing; or
  • retire a low-value service without replacing it.

Each is an active choice. A board can evaluate it more seriously than an undefined plan to wait.

Also specify the period. A three-month delay while a dependency is resolved has a different cost from repeated annual deferral. Some costs recur, some escalate, and some arise only if a particular event occurs. Putting them all into one confident total would create a new form of false precision.

Five places to look for the cost

The five categories below are prompts for investigation. They are not a promise that every deferred investment produces costs in each area.

1. Direct operating cost

Collect the expenses that keep the existing service available: licences, hosting, support, maintenance, specialist suppliers, incident work and internal technical time.

Separate activity that preserves operation from work that creates a new capability. The distinction will never be perfect. A security update may maintain the service and reduce future risk. Record the allocation method instead of pretending the categories are exact.

Look for costs outside the obvious technology budget. A content team may pay an agency to make routine changes. Operations may reconcile data manually because systems do not exchange it reliably. Finance may retain an older reporting process alongside the official one. These are part of the operating model the status quo requires.

2. Commercial opportunity

Lost revenue is tempting to exaggerate because it can make almost any investment look attractive. Build the estimate from first-party evidence.

Useful sources include enquiry and conversion data, losses where buyers explained the decision, abandoned journeys, client research, service complaints and the time between an initial request and a completed transaction. Distinguish between pipeline influenced by the digital experience and revenue that can reasonably be attributed to it.

Avoid inserting a generic sector conversion benchmark unless the measure, audience and acquisition mix genuinely match. If the firm cannot establish a reliable financial value, report the observed commercial friction and the evidence still needed. Unknown does not mean zero, and it does not justify an invented number.

3. Risk and continuity

An older platform may affect security, resilience, privacy, accessibility, records management, vendor support or regulatory compliance. Relevant specialists should assess the exposure and controls.

Probability-weighted financial estimates can help compare risks, provided the probabilities and impacts have a credible basis. A percentage chosen to complete a spreadsheet merely disguises uncertainty. Where evidence is limited, use scenarios: describe the event, plausible consequences, existing controls, early warnings and management response.

End-of-life dates, unsupported components, recovery-test results and incident history are stronger inputs than broad statements about legacy technology. Modern platforms also carry risk, especially during migration. Include transition risk on the investment side of the comparison.

4. Competitive position

Competitive deterioration is difficult to isolate. Firms should resist turning a subjective impression that competitors “look ahead” into a precise annual loss.

Ask the business development team what buyers raise during selection. Review pitch feedback and lost opportunities. Compare specific services and journeys rather than visual novelty. A competitor offering faster onboarding, clearer access to information or a useful digital service may create a meaningful gap. A fashionable interface by itself may not.

The cost may first appear as additional effort: more partner reassurance, bespoke workarounds, discounts or a longer sales cycle. These signals can support a decision even when causal revenue attribution remains uncertain.

5. People and capability

Poor tools can consume time, restrict learning and frustrate employees. The effect should be investigated rather than assumed.

Use task observation, service tickets, employee research, recruitment feedback and exit evidence. Estimate time spent on repeated workarounds only after checking a representative sample. Convert time to money carefully; an hour released from an inefficient task does not automatically become an hour of billable revenue.

Capability loss can also compound. When the organisation continually patches an old platform, its team may spend less time developing the skills required for a future operating model. That is strategically relevant even if it cannot be reduced to a payroll figure.

Build ranges that expose the assumptions

For each material category, record:

  • the observed evidence;
  • the calculation or scenario;
  • a low, central and high estimate where quantification is justified;
  • the assumption that changes the result most;
  • the confidence level;
  • the owner who can validate it; and
  • the date on which the estimate should be reviewed.

Use the same discipline for the proposed investment. Include internal time, content migration, training, change management, dual running, contingency and ongoing operation. Show which benefits depend on adoption or a change elsewhere in the business.

Do not add the high estimate from every category and present it as a likely outcome. Some risks overlap, some are mutually exclusive, and some commercial losses may already include operational effects. Finance and risk colleagues should challenge the model for double counting.

Where costs compound, explain the mechanism. A support contract may rise at renewal. Repeated workarounds may grow with transaction volume. A one-off missed opportunity does not automatically recur each year. The time profile should follow the evidence rather than a preference for a dramatic curve.

Compare more than two choices

A board paper framed as “approve the full programme or accept all these costs” can feel like advocacy. Include credible intermediate options.

For example, compare:

  1. maintain the current service with defined risk controls;
  2. address one urgent constraint while delaying wider replacement;
  3. run discovery or a bounded proof of concept;
  4. modernise in phases; and
  5. replace the platform as a single programme.

For each option, show cost, timing, organisational capacity, dependencies, principal risks, expected outcomes and the next irreversible commitment. The least expensive option may be appropriate if it preserves choices and addresses the immediate exposure. A phased route may also cost more overall while reducing transition risk. Those trade-offs deserve to be visible.

Turn deferral into a managed decision

If the board chooses to wait, record what was decided and why. Set conditions that would trigger reconsideration, such as a contract date, a support change, a risk threshold, a pattern in client evidence or a sustained operating cost. Name an owner and a review date.

This avoids the four-year deferral pattern described in the source article, where the same proposal returns without the accumulated cost ever being assembled. Costs spread across maintenance, support, recruitment, client service and lost time are easy to attribute to separate causes. A decision record makes the shared driver easier to see.

The purpose of cost-of-inaction analysis is not to manufacture approval. Sometimes the evidence supports waiting, reducing scope or stopping an idea. Its value is a fair comparison. The investment should not have to compete against an imaginary £0 alternative.

The Replatform Reckoning and The Customer Experience Dividend explore the evidence behind legacy-platform and customer-experience costs. The downloadable cost-of-inaction calculator provides the five-category structure for a board-ready comparison. Treat its output as a decision model with visible assumptions, not as an automatic business case.