Mid-year IT budget cuts: protecting delivery capacity

Leaders routinely confuse budget reductions with delivery capacity. Headcount and accounting spend are imperfect proxies for actual throughput. Throughput depends heavily on bottleneck skills, stable teams, and platform reliability. Decision speed and the amount of work in progress constrain output. My operating rule, drawn from the portfolios I have run rather than from any general law of systems, is that cutting a scarce capability reduces output disproportionately. The mechanism is queueing at a constrained resource: work arrives at the constraint faster than the reduced capacity can clear it, the queue in front of it lengthens, and everything downstream waits. Companies budgeting differently shift 10 to 20 percent of capital toward higher-return opportunities. They treat the budget as a roadmap for executing strategy. Across-the-board reductions ignore value, constraints, and operational exposure. False savings simply shift cost or delay value. They create restart expense and increase technical debt. An even cut feels fair but produces a smaller business across the board. Organizations capture only 31 percent of expected revenue benefits from recent digital transformations. A major outage costs more than $100,000 for 57 percent of companies. One in five pays over $1 million. One documented example: a company paused a self-service analytics dashboard and redirected engineering hours toward customer onboarding, with a direct impact on conversion.

The argument: redesign the portfolio around value and flow

Use one triage matrix for spend and roadmap decisions

Executives should assess each cost against business value and urgency. Reversibility and delivery-capacity impact matter heavily. Operational risk, dependencies, and time to financial impact determine the true cost of a cut.

Action	Use when	Typical targets	Primary test
Stop	Business value is low and no funded outcome depends on the work	Low-value projects, duplicate tooling, avoidable demand, headcount or contractor assignments left without a funded outcome	No funded outcome is lost and no dependency breaks
Defer	The outcome still matters, only the timing is flexible, and the delay does not compound	Delivery timing; technical debt outside security, resilience and the delivery bottleneck	Transition and restart cost stays below the saving
Reshape	The outcome still matters but its current shape is too expensive	Project scope, service levels, sourcing and contractor mix	The minimum viable outcome survives the reshaping
Renegotiate	The capability is needed but the commercial terms are uncompetitive	Vendor contracts, cloud commitments, licences	The same service level costs less, the saving lands inside the fiscal window, and operational risk does not rise
Protect	Delivery-capacity impact or operational exposure is high	Bottleneck roles, platform reliability, security and resilience, critical technical debt	Removing it reduces throughput or raises risk faster than it reduces cost

Defer and Reshape are the pair that people blur together, so the split is fixed: delivery timing belongs to Defer and project scope belongs to Reshape, never to both. The rule is one line — if the date can move, defer it; if the date cannot move, reshape the scope. Renegotiate or exit vendors when services duplicate other capabilities, commercial terms are uncompetitive, or the supported work stops. Reduce cloud spend and licenses when capacity is idle, usage is avoidable, or tools overlap. Stop projects with low value or no funded outcome. Defer technical debt only when it is not tied to security, resilience, reliability, or a delivery bottleneck. A defensible roadmap connects every line item to a business outcome. It sequences spend so early wins fund later bets. It names the risks of not spending in concrete financial language. Four in five serious outages could be prevented with better management, process, or configuration.

Rebuild the roadmap around minimum viable outcomes

Replace project-by-project defense with outcome-based prioritization. Rank commitments by mandatory obligations and economic value. Consider customer urgency and dependency criticality. Evaluate confidence and capacity consumed at the bottleneck. Preserve minimum viable commitments. Remove low-value features and reduce parallel work. Make explicit which outcomes or dates will change. A practical customer impact test helps here: if a customer would not notice a stoppage within thirty days, the work belongs on the cut list.

Sequence savings from reversible to destructive

Start with avoidable demand and duplicate tools. Target unused licenses and idle cloud resources. Next, reshape sourcing and contractor mix. Adjust vendor terms and service levels. Reshape project scope and defer delivery timing, checking transition costs in both cases. Treat reductions to bottleneck roles as the last resort. Protect security, resilience, and critical technical debt. Reversibility matters more than the nominal line item. Pausing a future-investment bucket can set a pipeline back by four to six months. Distinguish accounting recognition from operational implementation. When in-year cash is required, sequence differently. Renegotiating commercial terms and stopping unfunded work release cash inside the fiscal window; reshaping scope and service levels releases part of it; deferring projects and pausing investment usually shift recognition into the next period rather than producing cash now. The ordering follows from three mechanics rather than from preference: notice periods and termination clauses decide how fast a contracted cost actually stops, committed spend already recognised cannot be recovered by pausing the work, and deferral moves an obligation between accounting periods without removing it. Reject any action whose transition or restart cost lands in the same fiscal year and erases the saving.

Set guardrails before approving cuts

[spend-roadmap-triage-matrix]

Six protections are non-negotiable before any cut is approved. This list extends the Protect row of the triage table with two flow protections — team stability and work-in-progress limits — and it is the six-item list, not the table row, that governs approvals. A protection only becomes a guardrail when it is paired with the indicator that evidences it and the threshold at which it blocks an approved cut. The numbers below are the working defaults I start from in my own portfolios, not industry constants: calibrate each one against your organisation’s own baseline for the last four quarters before you let it block anything. – Bottleneck coverage — indicator: utilization at the constraint and the number of people qualified to work there. Blocks the cut when sustained utilization passes 85 percent, or when coverage would drop below two qualified people. – Team stability — indicator: voluntary churn on the delivering team. Blocks the cut when rolling three-month churn passes 10 percent annualised, or when one team would lose more than a quarter of its members in a single quarter. – Platform reliability — indicator: incident pressure against the quarterly error budget. Blocks the cut when the error budget is already spent, or when severity-one volume is rising month over month. – Security and resilience — indicator: critical findings past their remediation deadline and the date of the last successful recovery test. Blocks the cut when any critical finding is overdue, or when no recovery test has passed in six months. – Critical technical debt — indicator: unplanned work as a share of delivered work. Blocks the cut when unplanned work passes 20 percent for two consecutive months. – Work-in-progress limits — indicator: concurrent initiatives per team and delivery predictability. Blocks the cut when it would push a team above its agreed work-in-progress limit, or when predictability falls below 80 percent of committed items. The CEO, CFO, and CIO review savings realization together against these same indicators. Financial gains should not conceal delivery deterioration.

The counterarguments: why blunt cuts can look rational

“Equal percentage cuts are faster and fairer”

[reversible-cuts-sequence]

A uniform target is simple to communicate. It limits political bargaining. It creates rapid financial accountability. Nearly half of corporate executives are considering a budget cut for their function. Equal treatment of unequal capabilities is not economic fairness. It rewards bloated areas. It disables already constrained teams. Set a common savings obligation instead. Let portfolio evidence determine where the reduction lands.

“Headcount and contractors deliver the quickest savings”

Labor and contingent workforce costs are visible. They are sizable and reducible within the fiscal window. Cut duplicated management or persistently low-value capacity. Remove obsolete capabilities or roles left without funded outcomes. Cut contractors used for commodity augmentation, weak knowledge transfer, excess rates, or work that has stopped. Protect scarce expertise and transition knowledge. Without a defensible roadmap, hidden risks survive while visible costs get trimmed.

“Every project can absorb a little less scope”

Distributed scope reduction preserves stakeholder coalitions. It avoids visible cancellations. Thinly funding too many initiatives increases work in progress. It raises coordination costs and extends time to value. Finish fewer high-value outcomes with viable teams. One documented account shows the risk: venture-backed fashion firms cut about 40 percent of employees in March 2020 and reversed that judgement three months later.

My position: make savings conditional on preserved throughput

Approve scenarios, not isolated cuts

Protect delivery dates with less scope. Preserve scope with later dates. Accept defined risk for deeper savings. Show financial timing and affected outcomes for each scenario. Detail bottleneck impact and operational exposure. Map reversibility, dependencies, and the cost of restarting. The board should choose a portfolio outcome. Leading finance organizations treat budgets as plans designed for multiple possible outcomes. They use rolling scenario reviews to revisit assumptions.

Count savings only when value and risk are visible

Count savings when spend leaves the system without replacement by hidden internal effort. Exclude actions that cause later remediation, delay costs, transition expense, or elevated operational exposure. Require named owners and finance validation. Mandate delivery-capacity checks and risk acceptance for every decision. Six in ten executives find it difficult to quantify the benefits of individual technology investments, which is exactly why an unverified saving is worth less than it looks on the slide. Protect the capabilities that govern throughput. Remove scope, demand, and avoidable run cost until the portfolio fits the ceiling.

The call to action: reset the portfolio in 30 days

Days 1–10: establish the facts and freeze new demand

Form a CEO-CFO-CIO decision group. Confirm the in-year cash and accounting requirement. Assign one source of truth for spend and commitments. Map capacity, dependencies, and risks. Pause new discretionary demand immediately. Map bottleneck capabilities and identify the operational foundations that stay funded. Classify current spend using the triage matrix.

Days 11–20: build scenarios and make portfolio choices

This window produces exactly three scenarios, and each one carries a single named owner. The CIO owns the date-protected scenario, which holds every committed delivery date and takes the saving out of scope. The CFO owns the savings-maximising scenario, which meets the full cash target and states plainly which dates move. The CEO owns the middle scenario, which protects the two outcomes the board refuses to lose and accepts defined risk everywhere else. The artifact due on day 20 is one decision paper per scenario, no longer than a page: the saving it delivers and when the cash lands, the delivery capacity it consumes at the bottleneck, the outcomes and dates it changes, and the risks its owner is asking the board to accept. Nothing reaches the day-21 approval without all three papers on the table side by side.

Days 21–30: commit, communicate, and monitor

Approve owners and financial milestones. Lock in capacity guardrails and risk acceptances. Define escalation triggers for the chosen scenario. Communicate what will stop and what remains protected. A transparent conversation about cuts builds trust for subsequent decisions. Begin a joint governance cadence. Review savings alongside delivery and operational indicators. Use a rolling forecast to detect changing conditions and trigger a response.

[thirty-day-executive-plan]