Cost of Delay

Quantify what it costs not to have something yet, then prioritise by value lost per unit of time.

Every week of waiting loses the value the work would have delivered in that week.

To an executive

Every week a product is not live costs real money. Putting a figure on that loss lets leaders rank work by value rather than opinion or seniority.

To an engineer

Estimate the value lost per week for each item, then divide by duration. Small, urgent work often goes ahead of big work, and fixed deadlines change the order.

The idea

Don Reinertsen, writing about product development economics, argued that if you quantify only one thing, quantify the cost of delay: how much value is lost for each unit of time that something is not available. It turns arguments about priority, which are usually about opinion and seniority, into arguments about money and time.

Cost of delay on its own favours big items. Dividing it by the expected duration gives CD3 (cost of delay divided by duration), a ranking that maximises the value delivered over time. It often puts small, urgent pieces of work ahead of large, impressive ones.

The shape of the cost matters as much as its size. Some delays cost the same every week; others rise sharply, or fall off a cliff at a fixed date such as a peak season, a licence renewal or a regulatory deadline.

Why it works

It gives everyone the same unit. Features, platform work, risk reduction and compliance can be compared once each is expressed as value lost per week, instead of each side arguing in its own terms.

It makes waiting visible. Queues of approved but unstarted work look free on a roadmap. Putting a price on them shows that a long backlog has a real, ongoing cost.

Rough numbers are enough. Most priority calls change with an order-of-magnitude estimate. Being explicit about the assumptions does more good than being precise.

Where it applies

Portfolio prioritisation, where architectural and platform work must compete with product features for the same funding.
Roadmap sequencing, deciding which capability or migration wave goes first.
Vendor and contract timing, where a renewal date or end-of-support date makes the cost of waiting jump.
Making the case for architecture work, by stating the cost of a missing capability, such as a checkout that cannot scale before peak season, in business terms.

How to apply

State what the work unlocks: revenue, cost avoided, risk reduced or options created.
Estimate the weekly cost of not having it, in money where you can, and agree the assumptions with the business owner.
Describe the urgency profile: steady, rising, or tied to a fixed date.
Divide by duration to get CD3, and rank the portfolio by it.
Revisit the ranking when assumptions change, such as a new deadline or a slipped dependency.

Example scenario

An enterprise platform migration was deprioritised in every planning round, because it was framed as "technical clean-up" next to visible product features.

The architects reframed it: each month of delay meant paying for duplicate licences on the old platform and held back two product launches that depended on it. Expressed as a cost per month, and set against its duration, the migration ranked near the top of the portfolio and was funded in the next quarter.

Questions to ask

What does it cost the business, per week, not to have this yet?
Is that cost steady, rising, or tied to a fixed date such as a contract renewal or regulation?
Ranked by cost of delay divided by duration, where does this sit against the rest of the portfolio?

When not to use it

Work whose value cannot sensibly be priced, such as legal obligations or safety fixes, which should be done regardless of ranking.
Very small backlogs, where the overhead of estimating outweighs the benefit of a better order.

How it misleads you

False precision. A spreadsheet of cost-of-delay figures to the nearest euro looks authoritative but rests on guesses, and teams can game the inputs to move their work up. Treat the numbers as rough, show the assumptions, and use them to start the conversation rather than end it.