The Cost of Not Deciding

Why this matters

Not deciding feels like keeping your options open. Usually it is spending them. A hire you are "still thinking about" is paid for every week in overtime, turned-away calls and a crew that can see you are stuck. A truck you have not decided to replace keeps breaking on the schedule it chooses rather than the one you would have. The owner rarely sees this as a cost because no invoice arrives for it, and because a bad decision feels like your fault while a decision you never made feels like circumstance. This card names what an undecided question actually costs, works one delayed hire through to the end, and gives a simple test for telling a sensible wait from a drift.

What an open loop is

An open loop is a decision that is known to be needed, has not been made, and has no date by which it will be. The last clause is the one that matters. "I'll decide on the hire after I see March's numbers" is not an open loop; it is a deferral with a condition. "I need to think about the hire" with no end is an open loop.

Psychologists call the habit of judging harm from inaction more gently than harm from action omission bias (studied by Jonathan Baron and colleagues from the early 1990s). If you hire and it goes badly, you did that. If you do not hire and the shop bleeds overtime for a quarter, it feels like the market did that. Same owner, same outcome, different sense of blame, and the asymmetry pushes toward waiting.

Three kinds of cost

Hours. Every open loop gets re-opened. You re-run the numbers on the drive home, you discuss it again with the office manager, the lead tech asks again. None of that moves the decision; it re-litigates it. Owners who track it are usually surprised: two hours a week on one undecided question is common.

Morale. People can see a stuck decision, and they adjust. A crew working overtime while the owner "thinks about" help draws a conclusion about how much the owner values their evenings. An office manager whose proposal sits for six weeks stops making proposals. This cost is slow and it does not reverse when the decision is finally made.

Options. This is the one owners miss. Waiting does not hold the options still; some of them close. The candidate who applied takes another offer. The supplier's quote expires. The slow season, when a new system is cheapest to roll out, passes. A decision made late is frequently a choice among worse options than the ones on the table at the start.

Worked example: a hire delayed fourteen weeks

The shop: five techs, an office of two, a booking lead time that normally runs 2 working days. Demand rose in late winter and the lead time crept to 6 working days. The owner started weighing a sixth tech in week 1 and made the hire in week 14. Here is what the fourteen weeks held.

WEEK  WHAT HAPPENED
----  -------------------------------------------------
 1    Lead time 6 days. Owner: "let me think about a hire"
 2    Strong candidate applies. Owner asks for a week
 3    Dispatcher starts logging pushed or declined calls
 5    Candidate takes a job at another shop
 6    Twelfth straight week of rising calls. Trend clear
 9    Owner re-runs the numbers, still undecided
11    Senior tech gives notice, cites the overtime
14    Owner hires. Weaker candidate. Headcount: 5

The overtime premium. Across the fourteen weeks the five techs averaged 6 overtime hours each per week, 30 overtime hours a week, 420 overtime hours in total. The base pay on those hours would have been spent on someone either way. The extra half on each overtime hour was the pure cost of buying capacity from tired people instead of hiring it: 420 times 0.5 is 210 straight-time hours, a little over five weeks of one tech's 40-hour pay (210 divided by 40 is 5.25).

The lost work. From week 3 the dispatcher logged an average of 9 calls a week that were declined or pushed past 5 working days. Over the 12 logged weeks (weeks 3 through 14) that is 108 calls. The office's own estimate, from calling a sample back, was that about 40 percent went to another company: roughly 43 jobs. Those were not just one job each; a share of them were new customers who will now call the other shop first.

The owner's hours. The owner estimated 2 hours a week thinking about, discussing and recalculating the hire: 28 hours over fourteen weeks, all of it spent on a question whose answer did not change after week 6.

The options. The week-2 candidate was the best the shop saw all spring. The hire in week 14 was the best of who was left. And the senior tech's notice in week 11 meant the week-14 hire did not grow the shop; it replaced him. After fourteen weeks of deciding, the shop had the same five techs it started with, one of them new and one experienced tech gone.

Did waiting buy information? Partly, and that is the honest defense of any wait. In week 1 the rise in calls could have been a two-week spike. By week 6 the dispatcher's log and twelve weeks of call counts had settled it. So the information worth waiting for arrived around week 6, and weeks 7 through 14 cost the full rate and bought nothing. A decide-by date of week 6, with "is the call rise still there?" as the question, would have closed the loop eight weeks earlier and probably kept both the candidate and the senior tech.

Deferring versus drifting

A wait is a legitimate decision when it has three things written down:

• A date. "We decide on the hire by the end of week 6."
• The question the wait is for. "Is the call volume still up after the spike period?"
• What answer would change the decision. "If weekly calls are back within 10 percent of last year's level, no hire; if not, hire."

If you cannot write the second line, there is nothing the wait will tell you, and you are drifting. If you cannot write the third, the wait will tell you something and you will not act on it. Writing these three lines takes about two minutes and it converts an open loop into a scheduled decision. When the date comes, decide, even if the answer is "defer again with a new date and a new question." A second deferral with a reason is fine. A third on the same question is a sign the question is not really what is holding you.

When waiting is genuinely right

When the information is coming and it matters. A pending contract, a lease renewal, a supplier's price letter: if a specific fact is due on a specific day and it could flip the answer, wait for it. The test is whether you can name the fact.

When the decision is cheap to make later and expensive to undo. Signing a second lease or buying a building are worth waiting on if nothing is being lost meanwhile. Whether a decision can be cheaply reversed is its own subject; see related: Reversible and Irreversible Decisions in a Small Shop. The trap is treating a reversible decision (a hire with a probation period, a three-month software trial) as if it were permanent, and waiting on it as if it were.

When you are in no state to decide. Exhausted or angry is a good reason for a 24-hour pause, not a 14-week one. See related: Making Big Decisions When You Are Exhausted or Angry.

Finding the open loops in your own shop

Most owners carry three to six open loops at any time without listing them. Write them down: every decision you know is needed and have not made. Next to each, put the week it first came up. Anything older than four weeks with no decide-by date gets one of three treatments that day: decide it, schedule it with the three lines above, or drop it explicitly ("we are not opening a second location this year") and tell the people who were waiting.

The third option is underused. Many open loops are decisions the owner has already made privately (not this year) and never said out loud, so the team keeps raising them and the owner keeps re-thinking them. Saying no closes the loop as surely as saying yes.

What the list usually shows: the oldest loops are people decisions (a tech who is not working out, a hire, a partner conversation) and the newest are equipment and software. People decisions drift longest because they are the most uncomfortable to make, which is also why they accrue the largest morale cost while they wait.

The failure in the other direction

None of this argues for deciding fast in general. An owner who decides every question the moment it is asked, with no time for the information that is actually on its way, makes avoidable mistakes and then reverses them, which costs the crew trust in a different way. The aim is not speed. It is that every decision has either an answer or a date, and that the date is set by when the information arrives, not by when the discomfort fades.

References

• Spranca, Minsk and Baron, research on omission bias in judgments of harm (Journal of Experimental Social Psychology, 1991)
• See related: Reversible and Irreversible Decisions in a Small Shop
• See related: Making a Decision With Missing Information
• See related: Status Quo Bias With Software and Suppliers
• See related: The Cost of Being Your Own Bottleneck