Skip to content
Ashish Kumar AgnihotriOperating Partner

The vetting guide

25 Questions to Ask a Fractional COO Before You Hire

The right questions make a fractional COO interview impossible to bluff. These twenty-five are organised into five themes — scale experience, operating model, measurement, transfer and commercials — and each comes with the reason it works and what good and bad answers sound like. Ask them in order; the pattern across the answers is the real result.

Fractional COO interviews fail in a predictable way: the buyer asks about philosophy and chemistry, the candidate answers with fluency, and everyone leaves impressed with no evidence exchanged. Fluency is the cheapest thing a senior advisor owns. The questions on this page are built to be expensive to answer — they demand specifics only lived operating experience produces: numbers the candidate was accountable for, artefacts that survived their exit, engagements that ended well. A candidate who has genuinely run operations finds these questions easy and rather enjoyable. A candidate who has been near operations finds them uncomfortable, and the discomfort is the data.

The five themes mirror the anatomy of a good engagement. Scale experience establishes whether the judgement you are renting was actually built at the scale you are heading toward. Operating model tests whether they can show you, concretely, what they would install in your business. Measurement establishes how both of you will know it is working. Transfer probes whether your team ends up holding the model or holding a dependency. Commercials reveal how they think about accountability — pricing structure is a character disclosure, not a procurement detail. Weakness in any one theme is survivable; weakness in two or more is a forecast.

I have spent nineteen years on the answering side of tables like this — most recently as Senior Director, Business Excellence at Publicis Groupe, responsible for quality and delivery across 500+ clients and 2,000+ teams — so these are the questions I believe a serious operator should be able to answer without preparation, and the ones I would ask before hiring anyone into an operating seat, including a competitor. Use them with the one-page mandate described in the hiring guide: questions work best when every candidate is answering against the same real problem. Consistency of the brief is what turns a set of impressions into a genuine comparison.

In depth

What you need to know.

How to run the interview with these questions

Send your one-page mandate — the two or three operating problems with current numbers — before the conversation, so every answer lands against your reality rather than a generic company. Then work the themes in order: scale first, because it filters hardest; commercials last, because price only means something once you know what you would be buying. You will not need all twenty-five questions with every candidate — a strong operator settles the early ones quickly, and a weak one rarely survives the first ten. Take notes on specifics, not impressions: every number, artefact and named mechanism goes in one column, every adjective and framework in another. At the end, the ratio of the first column to the second is a better hiring signal than anything you felt in the room.

Scale experience: questions 1–5

Ask: What is the largest operation you have personally run — team size, volume, budget? Which numbers were you accountable for, and where did they end up? Walk me through one system you built that still ran after you left. What broke at scale that you did not expect, and what did you do? And: is our size big, small or familiar territory for you? Good answers are dense with specifics — an operator who ran quality across 2,000+ campaigns can tell you the score moved from 95% to 99% and exactly which checks moved it. Bad answers hide inside logo lists and the word “we”: prestigious employers, no personal accountability. The fifth question is a trap worth setting — candidates who inflate to impress will call your scale familiar when their history says otherwise.

Operating model: questions 6–10

Ask: Against our mandate, what would your first ninety days look like? What artefacts would exist at day thirty, sixty and ninety? What would you deliberately not change in the first quarter? How do you decide what gets fixed first? And: describe the weekly operating cadence you would run here — who attends, what is on the table, what leaves decided. These questions test whether the candidate installs systems or attends meetings. Good answers sequence diagnosis before prescription, name concrete artefacts — a scorecard on numbers you already trust, an operating cadence, ownership mapped to roles — and show restraint about what waits. Bad answers arrive with the solution already decided before anyone has looked at your data, or dissolve into methodology names and the promise of a framework. Certainty before diagnosis is theatre.

Measurement: questions 11–15

Ask: How will we both know this is working by day ninety? Which two or three numbers would you put on our scorecard, and why those? How do you baseline before changing anything? Tell me about an engagement where the numbers did not move — what did you do? And: how do you separate your impact from everything else that changed at the same time? Strong operators welcome this theme; it is where they live. Good answers prefer numbers your business already trusts — cycle time from the finance system beats a new dashboard — and treat attribution honestly, including the limits of it. The fourteenth question is the revealing one: everyone has had a number refuse to move, and a candidate with no such story is editing. Bad answers offer activity as evidence: workshops run, meetings held, decks produced.

Transfer: questions 16–20

Ask: How does an engagement of yours end? What did your last client keep after you left — show me the shape of it. Who runs the operating cadence when you are not in the room? How do you make yourself less necessary month by month? And: what happens to playbooks and documentation — in what format, owned by whom? These questions expose the difference between an operator who installs and hands over, and an advisor who becomes load-bearing. Good answers describe a deliberate taper: the client team chairing reviews by the second quarter, documentation in editable form on the client’s systems, an exit that was planned at entry. Bad answers cannot name a single artefact still alive at a former client, or quietly reframe permanence as loyalty. A dependency with a monthly invoice is not an operating model.

Commercial: questions 21–25

Ask: How do you price, and why that structure? What does your diagnostic cost, and what exactly does it produce? How many clients do you run at once, and where would we sit among them? What notice period do you work to, and what does exit look like contractually? And: when would you tell us not to hire you? Pricing structure is a disclosure of how the candidate thinks about accountability. Good answers: a fixed-fee diagnostic first, then a monthly retainer scaled to cadence and scope; a small number of concurrent clients, stated plainly; short notice, orderly transfer written in; and a genuine answer to the last question, because real operators decline wrong-fit work. Bad answers: hourly billing for an operating mandate, unlimited availability, vagueness about exit, and no conditions under which they would say no.

Answers that should worry you, whatever the question

Across all twenty-five questions, certain patterns should register as warnings regardless of where they surface. The pronoun problem: careers narrated entirely in “we,” with no result the candidate personally owned. The missing number: ten or twenty years of operations with nothing quantified anywhere. The instant prescription: your problems diagnosed and solved in the first meeting, before any data. The framework reflex: every question answered with a methodology name rather than a mechanism. The immortal engagement: no story of one ending well, artefacts intact, team self-sufficient. And the measurement dodge: discomfort with being judged on numbers, usually dressed as sophistication about attribution. One of these can be nerves or modesty. Two is a pattern. Three is a forecast of your own engagement, delivered free of charge in the interview.

Scoring the answers and making the decision

After each interview, score the five themes from one to three: three where the candidate produced verifiable specifics — numbers, artefacts, named mechanisms; two where answers were plausible but general; one where the theme was dodged or answered with adjectives. Weight scale experience and transfer highest: the first tells you whether the judgement is real, the second whether you keep anything after it leaves. Verify the top candidate’s best claims through references — ask referees about the exact systems and numbers cited, and listen for whether they saw them or heard about them. Then, rather than extending the interview process, buy the smallest real unit of work: a fixed-fee diagnostic. Two to four weeks of actual output against your actual operation settles what a sixth conversation never will.

Questions

Common questions.

“Walk me through one system you built that still ran after you left.” It compresses the whole interview: it demands real operating experience, a concrete artefact, a transfer that worked, and a result that survived without the person attached. Strong candidates answer with specifics — what the system was, what number it moved, who runs it now. Weak candidates retreat to projects they were near or frameworks they admire. If you have time for only one theme beyond it, ask how their engagements end; permanence dressed as partnership is the most expensive pattern in this market.

As a filter that tightens, not a script to complete. Send your one-page mandate in advance, open with scale experience, and let the candidate’s answers decide the pace — strong operators clear the early questions fast, weak ones rarely need the full list. Note specifics rather than impressions: numbers and artefacts in one column, adjectives and framework names in the other. Across sixty to ninety minutes you will comfortably cover the themes that matter. Save commercials for last, once you know what the money would actually buy.

Dense, specific and personally owned. A real operator names the operation’s size, the numbers they were accountable for, and the mechanism that moved them — a quality score lifted from 95% to 99% across 2,000+ campaigns and 450 clients, a billing cycle cut from roughly two months to fifteen days across 75 entities, a newsroom scaled fourfold to around 400 stories a day. The figures will differ; the texture will not. Vague answers lean on employer prestige and the word “we.” Scale you merely stood near does not transfer; scale you ran does.

Ask for the shape of one — a first-ninety-days walkthrough against your mandate: what they would examine, install and measure, and which artefacts would exist by day thirty, sixty and ninety. The specificity of that answer is the strongest interview signal available. But treat a detailed final plan demanded free as a bad trade for both sides: a serious operator will not prescribe before diagnosis, and any who will is guessing. The honest structure is a fixed-fee diagnostic — a small, priced unit of real work that replaces speculation with evidence.

The five in the transfer theme: how engagements end; what the last client kept; who runs the cadence when the operator is absent; how they make themselves less necessary each month; and who owns the documentation, in what format. Good answers describe a planned taper — your team chairing the weekly review within a couple of quarters, playbooks in editable form on your systems, an exit designed at entry. If a candidate cannot point to a single artefact still alive at a former client, what they sell is presence, and presence leaves with them.

Five: the pricing structure and the reasoning behind it; the diagnostic’s fixed cost and exact outputs; concurrent client load and where you would sit; notice period and what exit looks like contractually; and the conditions under which they would advise you not to hire them. The structure of the answers matters more than the amounts. Diagnostic-then-retainer signals accountability; hourly billing on an operating mandate rewards slowness; unlimited availability signals an empty practice; and a candidate with no version of “do not hire me” has never turned down wrong-fit work.

Careers narrated in “we” with nothing personally owned. No quantified result across a long track record. Solutions prescribed in the first meeting, before any diagnosis. Methodology names offered where mechanisms were asked for. No story of an engagement ending well, with artefacts intact and the team self-sufficient. Resistance to agreed measures, dressed as nuance about attribution. And hourly billing proposed for an ongoing operating seat. Treat one as a caution, two as a pattern, and three as the interview telling you exactly how your own engagement would end.

Score each theme one to three on verifiable specificity, and weight scale experience and transfer highest — the first proves the judgement is real, the second determines what you keep. The best fractional COO shows a consistent profile: numbers they personally moved at relevant scale, a ninety-day answer concrete enough to hold them to, engagements that demonstrably ended with the client holding the model, and commercials built around a diagnostic and a measured retainer. Confirm the top claims with references, then buy the diagnostic rather than another conversation — evidence beats a seventh interview.

The next step

A short conversation settles most of this — and a fixed-fee diagnostic settles the rest.

Book a consultationThe diagnostic

Keep exploring

How to hire a fractional COOThe full buyer’s guide these questions belong to.Fractional COO job description & scope of workWrite the mandate candidates answer against.The fractional COO agreementTurn good answers into enforceable clauses.The fixed-fee diagnosticThe evidence-based alternative to a sixth interview.Fractional COO costBenchmarks for judging the commercial answers.Fractional COO vs consultantThe distinction many candidates blur.