Decision debt: the interest you pay on decisions nobody made

By Meet Patel · 2026-10-03 · 6 min read

Take a 30-person software company (a hypothetical) that never decided its annual-plan discount. Some buyers are offered 10 percent and others 15, depending on who asks and which manager happens to be online. Six account executives each lose about two hours a week building custom quotes and waiting for an informal yes. Every Monday the leadership meeting spends twenty minutes on the same open question and closes with “let's circle back.”

Nobody in that company has made a bad decision. Nobody has made this one either, and the company pays for the gap every week.

I call this decision debt: the accumulated, recurring cost of decisions a team needs and has not made. The cost is real, it compounds, and it appears on no report.

The metaphor is older than the idea

Ward Cunningham introduced the debt metaphor for software in his 1992 OOPSLA experience report on the WyCash portfolio management system. He wrote that “shipping first time code is like going into debt,” that a little debt speeds development as long as it is repaid promptly, and that “every minute spent on not-quite-right code counts as interest on that debt.” The danger he named was debt that never gets repaid.

The structure carries over to decisions almost unchanged. The principal is the decision that needs making. The interest is the recurring cost of operating without it. Default happens when the gap gets filled by whoever acts first, so the company ends up living with a decision it never chose.

One difference matters. Technical debt at least lives in a repository, where an engineer can point at it. Decision debt lives in the gaps between people, so there is no file to open. People who join later inherit the workaround and never learn that a decision was missing, which is why this kind of debt survives so long.

It also helps to separate debt from deferral. Choosing to wait for the renewal numbers before setting the discount, with a date written down, is a loan with terms. Waiting with no date and no owner is debt. The two look identical in a meeting and cost very different amounts.

Three ways the interest is charged

Interest on an unmade decision arrives in three forms, and each has a symptom you can search for.

Workarounds. When the rule does not exist, people build a private one. In the discount example, each account executive keeps a personal sense of what is acceptable. The symptom is a process that exists in several slightly different versions, such as three spreadsheets that each calculate commission differently.

Duplicated effort. Two teams work on the same open question in parallel because neither knows the other is waiting on it. The symptom is two people presenting the same analysis in the same week, or two tools bought for one job.

Re-litigation. A topic returns to the agenda because it never closed. Each return costs the meeting time plus the preparation, and the preparation is usually the larger share. The symptom is a slide that was presented once and now comes back under a new title.

All three are measurable in hours, which is what makes the debt priceable. Measure your own company directly; I have not found a survey that isolates the cost of decisions that were never made.

Price one item before you argue about it

The formula is short.

Interest per week = (people × hours each) × loaded hourly cost

Use the loaded cost (salary plus overhead) that finance already uses. If nobody has one, pick a round placeholder and say it is a placeholder. Here is the discount decision with clearly illustrative figures, at $50 an hour:

• Workaround: 6 account executives × 2 hours = 12 hours a week.
• Re-litigation: 5 leaders × 20 minutes is about 1.7 hours of meeting time, plus about 1 hour of preparation by one person.
• Total: roughly 14.7 hours a week, or about $733 a week, or about $38,000 a year.

Now price the decision itself. One analyst spends four hours pulling discount history, and three people spend ninety minutes agreeing a rule. That is 8.5 hours, about $425 at the same rate.

The ratio of weekly interest to the cost of deciding is the number to sort by. Here it is a little under two, so the decision pays for itself in under a week. The example prices only time. It leaves out whatever margin the company loses to inconsistent discounts, and you should add that if you can measure it.

Take the inventory in an afternoon

You can do this with a handful of searches and one spreadsheet, using no new tool.

1. Agenda carry-overs. List every item that has appeared on a recurring meeting agenda two or more times without a recorded outcome.
2. Repeated questions. Search team chat for questions that recur, such as “which one do we use” and “who approves this.” Three repeats is enough to list it.
3. Manual patches. Ask each team lead one question: what are you doing by hand that you would stop doing if one thing were settled?
4. Parallel work. Ask two adjacent teams to list what they are waiting on. Any item that appears on both lists is duplicated effort.
5. Rewrite each item as a decision. Phrase it as a question with a name against it: “Which annual discount applies to which buyer? Owner: head of sales.” If you cannot name an owner, the missing owner is the first finding.
6. Price and rank. Estimate weekly interest and the cost to decide for each, and sort by the ratio.

I would expect a few items to carry most of the interest. That is useful, because it means you do not have to clear everything. I wrote earlier about why the decisions a company is not making are the expensive ones. The inventory is the practical version of that argument.

When waiting is the right call

Some open questions deserve to stay open. A team that has not yet seen the renewal numbers is right to hold the discount decision until they arrive. Three conditions make a deferral legitimate: information that will change the answer is genuinely on its way, the cost of waiting each week is small, and someone owns the trigger that ends the wait. If any of the three is missing, the deferral is debt with a polite name.

The second condition can be tested with the arithmetic above. If the weekly interest is $733 and the renewal numbers arrive in four weeks, waiting costs about $2,900. Whether that is a fair price for better information is a judgment the owner can now make explicitly and write down. Before the inventory, the same judgment was made by default, and nobody knew they were making it.

Pay it down by deadline, and borrow on purpose

Take the top three items from the ranking and give each one a decide-by date within two weeks, a single owner, and a named person who can overrule the owner. The decision does not have to be perfect. Many of these items are reversible, and a rule that is wrong in a known way is cheaper than no rule, because you can correct a rule that someone wrote down. Sorting decisions by how cheaply they can be reversed tells you which ones can be made this week.

Then change the terms on which you defer. The rule I would set is that no item appears on an agenda a third time without one of two things attached: a decision, or an explicit statement of what must become true before the team will decide, with a date. That converts open-ended debt into a loan with terms.

Write each outcome down where the next person will find it. A log of decisions stops the same question from coming back as new, and the template for keeping one takes ten minutes per entry. The same habit sits behind closing the gap between a signal and an action, because both depend on someone owning the next step.

A deferral always costs something, and the only open question is whether the terms are written down. Once a team knows what each unmade decision costs per week, it is harder to keep treating it as free.

Written by Meet Patel — startup operator and growth strategist in Dubai.

Read on themeetpatel.com