A forty-person company hired a fractional CTO in March. Two days a month, agreed on a call, confirmed in an email that ran to four lines. In July the board asked what had changed since March, and the chief executive had a list of meetings attended and no list of decisions made. The engagement was renewed anyway, because cancelling it would have been an admission.
That company didn’t buy a CTO. It bought a quantity. Sixteen hours a month of senior technical attention, invoiced monthly, with the title attached and the mandate left blank.
A fractional CTO is a senior engineering leader who runs your technical strategy part-time, usually on a retainer of one day a week or less. That’s the definition, and it’s the least useful sentence on this page, because every provider page we read agrees with it, and it leaves out the part that decides whether the money works. A fractional engagement is a budget. You spend it well or you waste it, and the difference is written down before the first call or it isn’t written down at all.
This piece treats it as exactly that. We count the hours you actually get, price the four shapes the engagement comes in, work out what the same money buys full-time, then spend the budget line by line: what to hand over first, what should never reach it, and what stays yours whatever you pay. Where we quote a price it is ours and published. Everything else is sourced.
What a Fractional CTO Is, in the Only Terms That Matter
A fractional CTO is a senior technology leader who owns your architecture, your hiring bar and your technical strategy on a part-time retainer, typically around eight hours a week. You’re not buying a person. You’re buying about thirty-five hours a month of senior judgement, a written mandate saying which decisions those hours cover, and an exit that hands the work back to you.
Three properties separate the real thing from what gets sold under the same name. Authority: a fractional CTO makes decisions rather than offering opinions, and somebody inside the company knows which ones. Continuity: they are there next month, which is what lets them refuse a decision they approved in a different form last quarter. A record: decisions land in writing, so the value survives the invoice.
Strip any of the three and you have something else. An advisor gives you an hour or so a month of pattern-matching against their own history, which is genuinely useful and is not this. A consultant delivers a report against a scope. A contractor delivers code. Only the advisor is reliably cheaper, since a scoped report or a full-time pair of hands can run past the retainer. If what you need is one of those, buy that instead.
One thing to know before you compare quotes. The market has no agreed definition of how many hours “fractional” means. Across the UK and US provider pages we read while writing this in September 2026, stated commitments ran from four hours a month to three days a week, each described as typical. That is roughly twenty-five to one on the same word. There is no standard, which makes the hours the negotiation rather than the label.
The Arithmetic Nobody Publishes: What Eight Hours a Week Actually Buys
Do the arithmetic weekly and the headline stops meaning what you assumed. Eight hours a week is about thirty-five hours a month, since a year runs 52 weeks rather than 48. Standing commitments take roughly five and a half of those eight hours: the leadership call, the architecture review, review of changes to the load-bearing parts of the system, writing the decisions down, the board report spread across the month, and reloading context between sessions. That leaves about two and a half hours a week, or eleven a month. Eleven is the real budget, and we have not seen a provider quote it.
Run the subtraction yourself before you sign anything, because the shape of the month decides what’s possible in it.
Where Twenty-Four of the Thirty-Five Hours Go
One hour a week goes to the leadership call, and that one is load-bearing: it is where technical reality meets commercial pressure in front of witnesses. One goes to the architecture review, where the expensive decisions get caught while they are still cheap. One goes to reviewing the changes that touch the architectural spine, and on a team shipping daily that is a floor rather than a ceiling.
One more goes to writing the decisions down, because a decision that isn’t recorded gets relitigated in six weeks by somebody who wasn’t in the room. Three quarters of an hour goes to context reload, the time before each session spent reading what moved since the last one. That is the tax on hiring part of a person rather than all of one. The board-ready report takes about three and a quarter hours a month, which is another three quarters of an hour a week.
That is five and a half hours a week committed out of eight, or about twenty-four a month out of thirty-five. Eleven are left. Now price the things you were hoping to spend them on. One engineering hire, done properly, is four to six hours across the scorecard, the loop and the debrief. A serious vendor evaluation is another four. One of each in the same month is eight to ten of the eleven, and the roadmap gets whatever is left.
Change the cadence and the numbers move. A fortnightly architecture review gives you back roughly two hours a month. Set the shape before the contract rather than discovering it in month three. None of this argues against fractional leadership. It argues against buying it without doing the arithmetic, which is what the company in March did.
The Hours the Engagement Costs You
Every hour of fractional CTO produces decisions that somebody inside your company has to carry out. The retainer is the smaller of the two numbers. A functioning engagement generates work for your lead engineer, your finance function and occasionally your legal counsel, and none of that appears on the invoice.
Budget an internal owner from week one. If nobody inside the company is accountable for implementing what the fractional CTO decides, you’re buying advice with a shelf life measured in weeks, and you’ll spend month four wondering why the architecture diagram is still a diagram.
Four Engagement Shapes, and What Each One Costs
Fractional CTO work comes in four shapes: a fixed-scope audit, a monthly retainer, advisory hours, and interim cover. Empyreal Infotech publishes two of them. An audit week runs at £8,000 for five days. A retainer runs at £5,000 to £8,000 a month for eight hours a week, cash only, invoiced monthly. Published ranges elsewhere are wider because there is no standard unit underneath them.
The fractional CTO cost question is usually asked as though there is one answer. There are four, and they buy different things.
A fixed-scope audit is the cheapest way to find out whether you need the other three. It is bounded, it produces artefacts, and it ends. Ours produces a diagram of the current architecture, six decision records on the most consequential choices already made, a ninety-day plan, and the list of risks that surface in investor diligence. If the audit says you do not need a retainer, that is a successful audit.
A monthly retainer buys continuity, which is the only thing that lets someone hold a position across quarters. It’s the right shape when decisions keep arriving rather than piling up once. It’s the wrong shape when you have one question.
Advisory hours buy access without authority. Useful, cheap, and easy to mistake for leadership when you are reading a proposal. If the contract has no mandate clause, you’ve bought advisory hours regardless of what the invoice says.
Interim cover is a different product. Full-time, temporary, usually covering a vacancy while you recruit, and priced accordingly. Buying interim when you meant fractional is an expensive sizing error, and the two words get used loosely enough that it happens.
One UK contractual point the provider pages we read do not mention. The off-payroll working rules apply where somebody supplies services through their own intermediary and would have been an employee if engaged directly. Responsibility for the status determination sits with the client in the public sector and in medium and large private-sector organisations, and with the worker’s own intermediary where the client is a small private-sector organisation, which covers most companies reading this. HMRC’s guidance on off-payroll working sets out who decides. Small is a statutory test built on turnover, balance sheet total and employee numbers rather than a feel for company size, the thresholds have moved recently, and unincorporated clients are tested differently again. Settle it with your accountant before the first invoice, not after the first audit.
Why a Day Rate and a Retainer Are Not the Same Purchase
A day rate buys presence. A retainer buys continuity. They can cost the same and still deliver opposite things.
Paying by the day makes every hour visible, which sounds like discipline and behaves like a disincentive. Nobody sends the Thursday message that prevents the Friday mistake if the Thursday message generates an invoice line. Retainers have the opposite failure: the money leaves whether or not anything happens, which is exactly why the mandate and the ninety-day measurable matter more on a retainer than on a day rate.
Pick the day rate when the work is episodic and you can name the days. Pick the retainer when you want somebody to notice things you didn’t ask about, and then hold them to a written outcome so that noticing things stays the job rather than becoming the deliverable.
The Full-Time Comparison, With the Awkward Bits Left In
This is where most comparisons flatter the fractional side, so here it is with the awkward bits left in. Our own figure for a full-time CTO base salary in London is £150,000 to £220,000. Employer National Insurance for 2026 to 2027 runs at 15% on earnings above a £5,000 secondary threshold a year, which adds £21,750 at the bottom of that band and £32,250 at the top. Call it £171,750 to £252,250 before pension, recruitment fees or equity. The Employment Allowance can cut an eligible employer’s total secondary bill by up to £10,500 a year, though a company already paying other staff will often have used it. The zero-rate categories for under-21s, apprentices and veterans are capped well below a salary at this level, so they will not touch this line. Run your own numbers rather than ours.
A retainer at £5,000 to £8,000 a month is £60,000 to £96,000 a year, or £68,000 to £104,000 in a first year that includes the audit week. On the ongoing-year figures, compared bottom against bottom and top against top, full-time costs roughly 2.6 to 2.9 times the retainer. Mix the ends and the spread runs from about 1.8 to 4.2 times, which is worth knowing before somebody quotes a single multiple back at you.
Now the part that sells nothing. Divide the full-time number by the hours it buys. About 225 working days of eight hours is 1,800 hours, so £171,750 is roughly £95 an hour and £252,250 is about £140. Our own page quotes £125 to £200 an hour, which counts the retainer month as forty hours. On the thirty-five hours eight hours a week actually produces it is £144 to £231, and that is the number to compare with. Against the bottom of the full-time band, fractional is dearer per hour. Against the top, the cheaper end of our range is roughly level.
So the saving isn’t a discount. It’s a decision to buy thirty-five hours a month instead of 1,800 a year, and both totals carry standing commitments the headline hides. It only pays if thirty-five is enough for the decisions in front of you. If it isn’t, you haven’t saved £100,000. You’ve underfunded the role and put a senior name on it.
One more thing the fractional side usually leaves out. Put a pension and a recruitment fee on the full-time column and the per-hour gap can close entirely. A £150,000 base with a ten per cent pension and a twenty-five per cent recruitment fee taken in year one reaches about £125 an hour, which is below the bottom of our own range on the same basis. The case for fractional is not that the hours are cheap. It is that you need fewer of them.
Spending the Budget: What to Delegate First
Delegate the decisions where being wrong is expensive and slow to reverse, and keep the ones that are cheap to undo. In practice that produces six categories, in this order: architecture, the hiring bar, build versus buy above a stated threshold, vendor commitments carrying a lock-in term, the security and compliance baseline, and the technical story you tell investors and acquirers.
- Architecture decisions that are expensive to unwind: the datastore, the boundaries between services, the auth model, anything that will still be true in three years.
- The hiring bar: the scorecard, the interview loop, and the standard for a yes. A hire below the bar is slow and expensive to unwind, and the code outlasts the person.
- Build versus buy above a named threshold: set the threshold in pounds or in sprints, then let that decision be theirs rather than a debate every time.
- Vendor and infrastructure commitments with lock-in: anything with a term, a migration cost or a data-gravity problem attached.
- The security and compliance baseline: what good looks like, what gets audited, and which findings block a release.
- The technical narrative for investors and buyers: the diagram, the decision record and the risk list that diligence will ask for anyway.
Then write the mandate down with a number in it. “Approves architecture changes” isn’t a mandate. “Approves any change that introduces a new datastore, adds a third-party dependency in the payment path, or costs more than one sprint of migration work” is a mandate, because you can tell from the outside whether it was followed.
Already paying for a retainer that hasn’t produced a written decision this quarter? Bring the invoice and the mandate to a thirty-minute call with Empyreal Infotech and we will tell you which of the two is missing.
What Should Never Reach the Budget
Sprint planning, ticket triage, standups, one-to-ones, on-call rotation, writing production features and chasing delivery dates. All of it is real work. None of it belongs in thirty-five hours a month, and a fractional CTO who takes it on will do it instead of the six categories above, not as well as them.
If the list of things you want covered is mostly that list, the honest answer is that you don’t need a fractional CTO. You need delivery leadership or delivery capacity, and the conversation is about MVP scoping and delivery or a permanent hire rather than a retainer. That answer costs us revenue and saves the relationship.
What You Cannot Delegate, Whatever You Pay
Five things never transfer: statutory director duties, budget authority, final hiring and firing decisions, formal acceptance of a named risk, and what the company says to a regulator or under a customer contract. A fractional CTO can recommend on all five, and should. Signing stays with you.
The first one is law rather than preference. Section 172 of the Companies Act 2006 requires a director to act in the way they consider, in good faith, would be most likely to promote the success of the company for the benefit of its members as a whole, having regard among other matters to the likely long-term consequences of a decision, and to five further factors the section lists. A fractional CTO is not normally appointed as a statutory director, and that duty does not move across with the invoice.
Do not read that as immunity, for you or for them. The Act defines a director as anyone occupying the position whatever they are called, treats a person whose instructions the board is accustomed to follow as a shadow director, and applies the general duties to shadow directors so far as they are capable of applying. It also carves out advice given in a professional capacity, which is where an external adviser’s argument usually starts, though that carve-out has been read narrowly. A mandate with real decision-making power sits closer to the line than a title suggests. It is worth twenty minutes with your solicitor when you draft it, and nothing in this article is legal or tax advice.
Risk acceptance is the one we see boards get wrong most often. The technical leader names the risk, prices it, and says what removing it would cost. Accepting it is a board act, because it’s a commercial decision about money and time. If your minutes say the fractional CTO accepted the risk, your minutes are wrong, and somebody will read them one day.
Budget authority is worth a sentence too. Give a spending ceiling rather than a veto: anything under the ceiling is theirs, anything over comes to you with a recommendation. That single line removes most of the friction that otherwise turns up in month three as a disagreement about scope.
The Column Everyone Leaves Out: Your Existing Engineering Lead
If you already have a lead engineer or an engineering manager, the fractional CTO’s mandate has to be written against theirs rather than over the top of it. In our experience the engagements that fail in companies with a working team don’t fail on competence. They fail because two people hold the same decision and one of them finds out in a meeting.
There are three honest ways to draw the line, and the only wrong answer is leaving it undrawn.
- Split by decision type: the fractional CTO owns architecture, the hiring bar and vendor commitments. Your lead owns delivery, the team and the week. Both report to you.
- Technical reporting line: your lead reports to the fractional CTO on technical decisions and to you on everything else. Cleanest on paper, hardest on people, and it needs saying out loud on day one.
- Succession shape: the fractional CTO holds no direct authority over your lead and is explicitly there to grow them into the role, with a date. Slowest, and the one that most often ends with you having a CTO you never had to recruit.
Pick one before the first call and tell the team which one you picked. The version where nobody says anything and everyone works it out from behaviour is the version where your best engineer starts interviewing elsewhere in month two, and you won’t hear about it until they resign.
Month Four: What a Failing Engagement Looks Like
Failing engagements don’t announce themselves, and month four is when the evidence is available but the sunk cost is still small. Five symptoms show up first, and any two of them together are enough to warrant a conversation rather than a renewal.
- No written decisions since the audit. The decision record has three entries and two of them were produced in week one.
- The weekly call has become a status update. Nobody brings a decision to it because decisions get made elsewhere and reported here.
- Engineers route around them. The team asks each other rather than asking the person you’re paying, which usually means the turnaround is slower than the deadline.
- The board report says “ongoing” in more than two rows. Ongoing is what a status looks like when there is no measurable underneath it.
- Nobody has said no to anything. Four months without a refused decision means either an unusually well-run company or an advisor who’s being agreeable for money.
There’s a single question that settles it faster than any of the five. Ask them to name the three decisions this engagement has made that would have gone differently without it. A good one answers in under a minute and the answers are specific. If it takes longer than that, you don’t have a fractional CTO. You have a subscription.
Ask it at month four, in writing, and put the same question in the contract as a scheduled review. The engagements that survive it get better because of it.
How You Know It Is Working: Five Numbers and a Baseline
Measure delivery rather than activity. DORA, the research programme run by Google Cloud, now publishes five software delivery metrics: change lead time, deployment frequency, failed deployment recovery time, change fail rate, and deployment rework rate. Take the baseline in week one, before anybody changes anything.
The baseline is the part that gets skipped, and skipping it is what turns measurement into marketing. Without a week-one number, every later reading is an assertion. With one, a mediocre improvement is still a fact you can act on, and a regression is visible while it is cheap.
Two of the five deserve particular attention, because they’re the ones a part-time leader can genuinely move. Change lead time responds to architecture and review discipline. Change fail rate responds to the standard for what gets merged. Deployment frequency tends to follow those two rather than lead them, so treat a frequency jump with no lead-time movement as a question rather than a win.
The Ninety-Day Measurable to Put in the Contract
Write one paragraph into the agreement that a stranger could audit. Something close to this: by day ninety, change lead time on the primary service is under an agreed figure taken from the week-one baseline, a decision record exists for every change that altered the architecture, the hiring loop has a written scorecard that’s been used on at least two candidates, and the risk register has named owners and dates against every entry.
Pick numbers you can measure on day one. A measurable you can’t take a reading for today is a sentence, not a commitment, and by month four it will have quietly become “ongoing”.
The Exit Is the Deliverable
A fractional CTO engagement that can’t end hasn’t worked. The exit artefacts are the product: a decision record somebody can read without the author present, a current architecture diagram, a hiring loop your team can run alone, and a handover to whoever holds the role next. If those four exist, the engagement paid for itself even if it ends early.
Median tenure across our 2025 retainers was fourteen months. Two of those retainers ended by handover to a full-time CTO, both at the client’s request, and in both cases we recommended the incoming hire in writing. That is the outcome the structure is built for rather than an accident it survived. We help interview the full-time hire, write the handover document and transfer the decision record. There’s no exit fee and no annual contract, and the thirty-day walk-away runs in both directions.
Ask any provider you’re considering what their median engagement length is and how the last two ended. It’s a harder question than it sounds, it takes ten seconds to answer honestly, and the shape of the hesitation tells you most of what you need to know.
When a Fractional CTO Is the Wrong Purchase
A fractional CTO is the wrong purchase in three common situations: you don’t yet know what you’re building, you need one decision examined rather than standing leadership, or you need evidence for a transaction rather than direction for a team. Each has a cheaper answer, and each of those answers is a smaller invoice than the one you were about to sign.
If the product isn’t defined, technical leadership has nothing to lead. The scope moves faster than the architecture can follow, and you’ll spend the retainer re-deciding things. What resolves it is a product discovery workshop that produces a defined thing to build, which then makes the architecture question answerable in an afternoon rather than a quarter.
If there’s one decision on the table and the rest of the year looks quiet, buy the decision. An independent architecture review is bounded, produces a written finding, and costs a fraction of twelve months of retainer. Retainers are for companies where decisions keep arriving.
If an investor or an acquirer is the reason you’re asking, what you need is an evidence pack rather than a leader. Technical due diligence produces the artefacts a buyer’s team will ask for, on their timetable. Hiring a fractional CTO six weeks before diligence to fix what diligence will find is a well-travelled and expensive route.
We’ve given all three of those answers to companies who arrived ready to sign a retainer. It costs us revenue. A retainer we talked someone out of is cheaper for both sides than one that renews out of embarrassment, which is what happened to the company in March.
How Empyreal Infotech Structures a Fractional CTO Engagement
Our fractional CTO engagement models are deliberately short and deliberately priced in public. It starts with a thirty-minute discovery call at no cost, built around fifteen questions on the state of your product, your team and your roadmap. If a retainer isn’t the right answer, we say so on the call.
If an audit makes sense, it’s £8,000 for five days and it produces four things: a diagram of the architecture as it actually is, six decision records on the most consequential choices already made, a ninety-day plan, and the risks that surface at Series A diligence. Plenty of companies stop there, which is a legitimate ending rather than a failed sale.
If a retainer makes sense, it’s £5,000 to £8,000 a month for eight hours a week: a weekly leadership call, review on every change that touches the architectural spine, decision records before code moves, hiring and interview support, and a monthly board-ready report. Delivered directly by Mohit, with no agency layer and no juniors behind the name on the contract.
It’s cash only. We don’t take equity, and the reason is worth saying plainly: equity aligns an advisor with an exit event rather than with your product decisions, and those two things point in different directions more often than anyone admits. Invoiced monthly, net fourteen days, thirty-day walk-away both ways, no annual contract and no exit fee.
The admission against interest, since this is the part of the article where we’re selling. At thirty-five hours a month we can’t run your delivery, we can’t be your escalation path at two in the morning, and we can’t replace a head of engineering. If that’s what the job needs, we’ll tell you on the discovery call, and you can book that call without a commitment attached to it.
FAQ: Fractional CTO Services
How much does a fractional CTO cost in the UK?
Empyreal Infotech publishes £5,000 to £8,000 a month for eight hours a week, plus £8,000 for a five-day audit week if you start with one. Our page quotes that as £125 to £200 an hour on a forty-hour month. On the thirty-five hours eight hours a week actually produces it is £144 to £231, which is the honest comparison figure. Ranges published elsewhere run much wider, because providers put very different hour commitments behind the same word.
What is the difference between a fractional CTO and an interim CTO?
The working distinction is hours and purpose. An interim CTO is full-time and temporary, brought in to cover a vacancy for a fixed period, and priced close to a full-time salary. A fractional CTO is part-time and ongoing, typically a day a week or less, and stays until the company outgrows the arrangement. Our own median across 2025 retainers was fourteen months.
How many hours a month does a fractional CTO actually work?
Eight hours a week is about thirty-five hours a month. On a working retainer, standing commitments take roughly five and a half of the eight hours each week, leaving about eleven discretionary hours a month for new problems. Ask any provider to break the week down before you sign, because the split between committed and discretionary time decides what the engagement can actually do.
Do fractional CTOs take equity instead of cash?
Some do, usually as part of a lower cash rate at early stage. We don’t. Equity ties the advice to an exit event rather than to your product decisions, and the two diverge more often than founders expect. Our retainers are cash only, invoiced monthly at net fourteen days.
What should a fractional CTO be responsible for, and what stays with me?
Delegate architecture, the hiring bar, build-versus-buy above a set threshold, vendor commitments with lock-in, the security baseline and the technical story for investors. Keep statutory director duties, budget authority, final hiring decisions, formal risk acceptance and anything the company says to a regulator or under contract. Write both lists down before the engagement starts.
What to Write Before You Take a Single Call
Return to that board meeting in July. The chief executive wasn’t badly advised and didn’t hire the wrong person. They bought a quantity without writing down what it was for, and four months later there was nothing to point at, because nothing had been agreed in a form anyone could check.
The fix costs an afternoon and no money. One page, written before you take a single call: which decisions move, what the spending ceiling is, who your existing lead reports to, what day ninety has to show, and what the exit looks like. Take that page to three providers and the conversations will sort themselves out, because the ones who can’t work inside it will tell you so.
If it helps to have someone read the page with you, the thirty-minute discovery call is free and ends with a recommendation rather than a proposal. You can book it here, and if the answer is something smaller than a retainer we will say so.
Write the mandate. Count the hours. Set the exit.