Most prioritization conversations ask one question: what is this worth if we build it? They rarely ask the other half of the same question: what is it costing us, every week, that we haven't built it yet.
That second cost is just as real as the first. It's just harder to see, because it never shows up as a line item—it shows up as revenue, market position, and opportunity that quietly never happened.
Cost of Delay: Putting a Number on Waiting
Donald Reinertsen calls this quantity Cost of Delay—the economic value of finishing something sooner rather than later, expressed in the same units as everything else that matters to the business: dollars per unit of time. He has described it as close to a master key for product economics, because once delay has a dollar figure attached to it, every other tradeoff—queue size, batch size, sequencing, staffing—can be evaluated in the same terms instead of argued about in the abstract.
That number gets turned into a sequencing decision through CD3: Cost of Delay divided by Duration. It answers a specific, practical question—given two initiatives, which one is more economically urgent to do first? The idea has since become the core mechanism behind Weighted Shortest Job First (WSJF), the prioritization formula used in the Scaled Agile Framework, where Cost of Delay (business value, time criticality, and risk reduction combined) is divided by job size to decide what gets sequenced next. Whatever the specific formula, the underlying move is the same: turn urgency into a number instead of an opinion, so it can compete on equal footing with every other number in the backlog.
Exhibit 1
Value sits ready to build
It waits in a queue behind other work
Each week of waiting has a real dollar cost
The cost compounds the longer it waits
By the time it ships, real value has already been lost
Why an orderly-looking backlog can still be quietly bleeding value every week.
"We Don't Have the Data" Isn't a Good Reason to Skip This
The most common objection to Cost of Delay is that it's hard to estimate precisely. Reinertsen's answer to that is direct: an imprecise economic estimate still beats no economic estimate. He points out that the typical gap between the best and worst items in a backlog, once you actually estimate their economic value, tends to run somewhere on the order of 50 to 1—meaning even a rough estimate is usually more than accurate enough to tell you which end of the backlog something belongs on. Waiting for a perfect number before you're willing to use one is not caution. It's a way of letting every item default to equal priority, which is itself a very expensive decision that nobody chose on purpose.
What This Looks Like When It's Real
This isn't just a framework on a whiteboard. Joshua Arnold and Özlem Yüce, who developed much of the practical Cost of Delay methodology now known as Black Swan Farming, documented its application at Maersk Line, one of the world's largest shipping companies, across a large enterprise product portfolio. One example from that work: a single feature had a Cost of Delay exceeding $200,000 per week. It sat in queue for 38 weeks. That queue time alone—work that wasn't being actively opposed or deprioritized, just waiting its turn behind other work—corresponded to nearly $8 million in lost revenue.
~$8M
in lost revenue — from a single feature sitting 38 weeks in queue at a $200,000-per-week Cost of Delay.
Source: Arnold, J. & Yüce, Ö., Black Swan Farming — Maersk Line case detail.
Nobody made an $8 million decision on that item. It simply sat in a queue, and the organization absorbed the cost without ever seeing it as a decision at all.
That is the core problem with cost of delay left unmeasured: it doesn't announce itself. It accumulates silently in exactly the items nobody is currently discussing, because attention naturally goes to what's active, not to what's waiting.
Why the Brain Doesn't Catch This on Its Own
There's a reason this particular blind spot is so persistent, and it isn't a discipline problem. Research on decision-making under delay has found that people are reasonably good at recognizing the opportunity cost of a large, distant decision—but systematically pay less attention to the same kind of cost when it's attached to something smaller or nearer-term. A backlog is made of exactly that: dozens of individually small, individually reasonable-looking delays, none of which trigger the mental alarm that a single big, obvious cost would. The $8 million at Maersk didn't arrive as one decision. It arrived as 38 individually unremarkable weeks.
Make the Invisible Cost Visible
None of this requires abandoning the frameworks an organization already uses. It requires adding one question to whatever process already exists: what does a week of delay on this actually cost, even roughly? Not a perfect number—a defensible one. Reinertsen's own point stands: the gap between your best guess and no estimate at all is far larger than the gap between your best guess and a perfect one.
The value of what you build is only half the prioritization equation. The other half is what it costs you, every week, to not have built it yet—and that half doesn't stop accruing just because nobody put it on a slide.
