What a Fractional COO Actually Does in the First 30 Days

You brought in a fractional COO because the business had outgrown the way you were running it. Three weeks later the org chart looks the same, nobody has rewritten a process, and you are quietly wondering what you are paying for.

That feeling ends more fractional engagements than poor work does. Usually the operator is not idle. Nobody agreed in advance what month one was supposed to produce, so the only available measure becomes how much visibly changed. In a well run month one, very little visibly changes.

A fractional COO's first 30 days do not buy change. They buy the mandate to change things in month two.

That sounds like an excuse. It is the opposite. An operator who reorganises your team in week one is guessing at your expense, because they cannot yet tell which of your problems are causes and which are symptoms. Four weeks spent mapping how work actually moves, naming the real constraint, and installing a rhythm the business can hold without them earns the right to move fast in month two.

This is a bigger market than the word fractional suggests. SMEs were 99.6 percent of Singapore enterprises and 69.2 percent of its employment on 2025 figures (Singapore Department of Statistics), and Australian small businesses employed 42 percent of the private sector workforce in 2022 to 2023 (ASBFEO). Almost none of them can justify a full time COO on a permanent salary. Many will need one for a day or two a week.

What does a fractional COO actually do in the first 30 days?

They climb a ladder of permission, one rung a week, and each rung has to hold before the next one is worth standing on.

We call it the Mandate Ladder. It exists because the failure mode of fractional work is not laziness, it is overreach. An experienced operator arrives carrying pattern recognition from three previous businesses and starts applying it before knowing which patterns transfer to yours.

Exhibit 1: The Mandate Ladder, week by week

Rung	Days	Mandate earned	Exit test you can run
1. Look	1 to 7	Permission to observe	The operator names a specific surprise
2. Name	8 to 14	Permission to diagnose	The constraint fits one sentence you could argue with
3. Run	15 to 21	Permission to change how work runs	A decision gets made that would previously have queued
4. Leave	22 to 30	Permission to change what the business does	Someone other than the operator chaired the review

Rem.Up framework. The exit tests are the point: each is observable by the client, not self reported by the operator.

Notice what is absent from all four rungs. No new org structure, no new tooling, no forty slide diagnostic. Those are month two decisions, and each gets cheaper once the ladder has been climbed.

Why does week one look like nothing is happening?

Because the most valuable thing an operator does in week one is refrain from acting, and restraint is invisible from the outside.

Three things go wrong when rung one gets skipped. Rem.Up field observation, from engagements with founder led firms in Singapore, Australia and Malaysia.

Restructuring before mapping. The operator arrives with a hypothesis from their last role and starts moving people. Reorganisations are expensive to reverse, and they spend the goodwill you need in month two.

Fixing the loudest problem. The loudest problem is rarely the binding one. It is usually a symptom, raised by whoever is most senior in the room.

Becoming the new bottleneck. Decisions that queued behind you now queue behind the operator. The queue feels shorter for a fortnight, then it is the same queue with a different name.

A good week one is unremarkable to watch. The operator sits in every recurring meeting without chairing one, follows the top three processes from the moment work arrives to the moment it is invoiced, interviews the people who do the work rather than only those who report on it, writes down where work waits and who it waits for, and publishes nothing.

There is a simple test at the end of week one. Ask what surprised them. A specific answer about how an approval or a handoff actually behaves means they were looking. “Broadly as expected” means they arrived with conclusions and spent the week confirming them.

Week one concentrates on decisions rather than processes because decisions are where the time goes. In Decision making in the age of urgency (McKinsey, 2019), surveying 1,259 managers and executives, respondents reported spending 37 percent of their working time on decision making, and 61 percent said most of that time was used ineffectively. Organisations that decide quickly were about twice as likely to make high quality decisions. Speed and quality travel together rather than trading off, which is why a good operator maps your decision queue before your process maps.

How do you find the real constraint instead of the loudest complaint?

You follow the queue rather than the noise. The binding constraint is wherever work waits on a person instead of on a process.

The company below is illustrative, built to show the arithmetic rather than to describe a real client. Assume a Singapore headquartered professional services firm: 55 people, USD 9 million revenue, a 12 percent operating margin, delivery split between Singapore and Kuala Lumpur. The founder's brief is that the project managers are not senior enough. Widely believed, and wrong.

Two weeks of mapping shows something else. Every scope change, rate exception and resourcing swap lands on the founder's desk, because nobody else holds the authority to approve one. The project managers are not junior. They are unauthorised.

The arithmetic, on assumptions we would state to a client exactly this way:

1. Four founder gated decisions per week across 46 working weeks: 184 decisions a year.

2. Median wait for a decision: 4 working days. That is 736 decision days of delay a year.

3. Not all waiting is idle, because people switch to other work. Assume each delayed decision holds up the equivalent of half a person day for every day it waits: 368 person days.

4. At a fully loaded USD 340 per person day, that is USD 125,120 a year.

5. Against operating profit of USD 1.08 million, being 12 percent of USD 9 million, that is roughly 11.6 percent of profit.

Illustrative, using the assumptions stated above. Change the switching assumption from half a person day to a whole one and the figure doubles. The number does not need to be precise. It needs to be large enough to justify a fix that costs nothing, and it is. It also excludes revenue that slips a milestone into the next quarter, which we leave out rather than model badly.

So the constraint gets written down as one sentence:

Delivery slips because scope and resourcing decisions can only be approved by the founder, and the founder is the busiest person in the business.

That sentence is testable against last quarter's misses, falsifiable so the team can argue with it, and it points at a fix that costs nothing: write down who can approve what, up to which threshold. Compare it with “we need better project management”. Vague constraints fund activity. Specific ones fund progress.

Two rules govern this rung. The constraint is agreed in writing, not implied on a call. And everything else found is published as a parked list with dates, so nothing looks quietly abandoned. Parking a problem in the open is a decision. Dropping it in silence is how trust leaks.

What does an operating cadence look like, and why install it before restructuring?

Because rhythm is cheaper than restructure, and it tells you whether a restructure is needed at all. More than once we have watched a business about to reorganise discover, four weeks into a functioning weekly review, that it had a visibility problem rather than a structure problem. Rem.Up field observation.

The cadence has three parts, and none require new software.

A weekly operating review. Fixed time, thirty minutes, standing agenda: what moved, what is blocked, who owns the unblock, what we decide today. If it runs long every week, either the agenda is wrong or the decision rights are missing.

Three numbers that predict rather than report. Last month's revenue reports. Proposals awaiting a decision, the average age of an open blocker, and the utilisation of whatever resource is scarce all predict. Choose the three that would have warned you about your last bad quarter in time to act.

Decision rights, written down. For each thing that keeps queueing, one line: who decides, up to what threshold, who gets told afterwards. The least glamorous artefact of the month, and reliably the one that moves the needle, because it is the only one that takes work off your desk permanently.

On where AI belongs, since it is the question we are asked most. It is good at the mechanical half: assembling the weekly pack, flagging items that have aged past a threshold, drafting the record. It is not good at the half that decides whether the cadence works, which is a person saying that a date is not real and being taken seriously. Automating a review nobody trusts produces the same failure faster. Install the rhythm first, then automate the admin around it.

How do you tell a deliberate operator from a slow one?

By whether artefacts appear on schedule. Deliberate operators produce something checkable at the end of every rung. Slow ones produce updates.

Six checks at day 30. If you cannot tick five, month one did not land, and the time to say so is before month two starts.

1. The constraint is written in one sentence you agree with.

2. A weekly review runs at a fixed time with named owners.

3. Three predictive numbers are reviewed as the same set each week.

4. Someone other than you chairs the review.

5. The cadence survived one full week without the operator in the room.

6. Month two scope is agreed, and it points at the constraint.

Checks four and five get skipped most often and matter most. A fractional COO who has made themselves indispensable by day 30 has failed, expensively. If every decision now routes through the operator instead of through you, you have not removed a bottleneck. You have rented a new one.

There is a real reason to hold this to 30 days rather than 90. In The First 90 Days (Harvard Business School Press, 2003), Michael D. Watkins defines the breakeven point as the moment a new leader has contributed as much value as they have consumed, and reports that 210 CEOs and presidents put the mean at 6.2 months for a midlevel manager. A fractional operator does not get six months. They are in one or two days a week and the clock is visible on every invoice.

Day 30 is the natural decision point. If the constraint turned out smaller than expected, scope down and pay for less. If structural, scope up deliberately rather than by drift. And if the operator is the wrong fit, month one is a far cheaper place to find out than month six. That optionality is the commercial logic of buying operating leadership by the fraction rather than the headcount.

What to do on Monday

Write your constraint in one sentence, before anyone else does. Specific enough that a colleague could disagree. Then compare it with what your operator writes in week two. If the two match word for word, one of you is not looking hard enough.

List the decisions that waited on you last month. Not the meetings, the decisions. Count them and note how long each sat. In most founder led businesses that list is the engagement, and it takes twenty minutes to produce.

Pick your three predictive numbers. Look back at the last quarter you got wrong and ask which numbers would have flagged it while you could still act. If you cannot find any, that is itself the finding.

Set the review before you set the structure. A thirty minute weekly meeting with named owners costs nothing and reveals more about where the business is stuck than a reorganisation will.

None of this needs a full time hire, and none of it needs a consulting engagement that ends in a report you implement alone. It needs a senior operator working inside the business, for the fraction of time the problem actually requires.

If you want a second view on where your constraint sits, we run a thirty minute conversation for exactly that. We either name it, or we tell you that you do not need us. Book a 30 minute operations call, or read more about how we work at rem-up.com.

Innovate. Optimize. Grow.