Topic 57 of 64· Software engineering and digital delivery connection
Cost of Delay in Public Programs (CoD)
Cost of Delay is the public value lost per unit time that a program, service, or system change is not yet delivered. It is the master bridge metric of this group of topics: it converts "the go-live slipped six months" into pounds per week, or into WELLBYs per week, so that delay can be argued about in the same currency as the business case itself.
Why it matters
Reinertsen's rule — "if you only quantify one thing, quantify the Cost of Delay" — travels into government almost unchanged, because public programs are unusually exposed to it: business cases are approved against a forecast benefit stream, but the stream only starts flowing at go-live, and every week of slippage is a week of forgone value that nobody prices on the risk register. The National Audit Office's repeated scrutiny of Universal Credit's rollout (see its "Rolling Out Universal Credit" reports, https://www.nao.org.uk/) illustrates the pattern: schedule slippage was tracked and reported, but the pounds-per-week cost of not yet delivering the reformed system to the next tranche of claimants was rarely stated as a headline figure, even though it is the number that should have driven prioritization and escalation. Without a CoD figure, a delayed program looks like a schedule problem for the delivery board; with one, it is a value-erosion problem for the accounting officer.
The math
CoD = benefit per unit time forgone while undelivered (£/week or WELLBYs/week)
Total delay loss = CoD × delay duration
Benefit streams to sum for public programs:
cash-releasing savings (fraud/error reduction, avoided temporary costs)
+ non-cash capacity freed (caseworker/officer hours × loaded cost)
+ wellbeing benefit (WELLBYs × £13,000/WELLBY, HMT Green Book
wellbeing supplementary guidance, 2019 prices)
For citizen-facing services, denominate in wellbeing as well as money — see wellbeing-adjusted life years for the underlying unit, and opportunity cost in public spending for what the delayed pound could otherwise have funded.
Worked example
Local authority: a housing-benefit system upgrade cuts overpayment error by £150/claim/year across 20,000 live claims.
Annual benefit = 150 × 20,000 = £3,000,000/year
CoD = 3,000,000 / 52 ≈ £57,700/week
A 12-month implementation delay costs 52 × 57,700 ≈ £3,000,000 in avoidable error.
Central government agency: a disability-benefit assessment service, delivered six months (26 weeks) later than planned, means 200,000 claimants/year wait an average three weeks longer for a decision. Each extra week of financial uncertainty is modeled as a −0.0018 WELLBY (life-satisfaction point) effect:
WELLBY loss per claimant = 3 × 0.0018 = 0.0054
Annual WELLBY loss = 200,000 × 0.0054 = 1,080 WELLBYs/year
CoD_wellbeing = 1,080 / 52 ≈ 20.8 WELLBYs/week
CoD_money = 20.8 × £13,000 ≈ £270,000/week of wellbeing value
A 26-week delay therefore "costs" roughly 540 WELLBYs — worth about £7 million at the Green Book's wellbeing valuation — reframing a missed go-live date as a citizen-welfare event, not a project-management footnote.
Software engineering connection
CoD is what makes DORA metrics and flow metrics financially legible: lead time in the pipeline × CoD is money (or wellbeing) burned in queues before it ever reaches a citizen. Concretely:
- Prioritization: rank a backlog by CoD ÷ duration rather than by stakeholder seniority — the software engineering analog of the Green Book's requirement to appraise options on value, not on who is asking.
- Procurement: a 12–18 month framework procurement cycle has a CoD; pricing it changes the urgency case for accelerated routes, and feeds directly into build vs buy decisions where time-to-value is a decision driver.
- Benefits case: every CoD figure quoted at approval should reappear at benefits realization — if the delay cost was real, the accelerated benefit should be measurable after go-live.
Pitfalls
- Assuming linear CoD: some public services have deadline-shaped value (a statutory compliance date — CoD jumps to enforcement-risk levels after the date, near zero before) rather than a smooth weekly rate. Classify the urgency profile before multiplying.
- CoD on outputs nobody needs: delay only has a cost if the undelivered thing has value; a system nobody will use has zero CoD regardless of how late it is.
- Double-counting delay and discounting: the social discount rate already prices time on multi-year appraisal horizons; CoD is the within-horizon, operational version for weeks and months. Use CoD for schedule slippage, NPV shift for multi-year re-phasing.
Sources
- Reinertsen DG, The Principles of Product Development Flow, Celeritas Publishing, 2009.
- HM Treasury, Green Book supplementary guidance: wellbeing. https://www.gov.uk/government/publications/green-book-supplementary-guidance-wellbeing
- National Audit Office, reports on Universal Credit rollout. https://www.nao.org.uk/