Fractional CTO or agency? The honest answer is that neither label decides it. A fractional Chief Technology Officer (CTO) sells direction; an agency sells delivery. What actually determines whether the engagement works is who is accountable for the result once it is built and running, not which of the two is cheaper or who bills the most hours. Both can be staffed by capable people. Both can give you sound advice. The framing feels like it captures the decision, and it hides the one fact that matters most: the label tells you almost nothing, and the accountability tells you almost everything.
We have written separately about what CTO as a service includes and what it costs. This piece is about the choice that comes just before that one, the one owners actually circle on the page: should the help be a senior individual who sets direction, or a team that builds, and which of those gets you to a working outcome you own. If the question is broader still, who should build the software at all, we compare a freelancer, an agency, and a firm.
What is a fractional CTO?
A fractional CTO is senior technical judgment, bought part-time. The deliverable is decisions, not code: what to build and what to buy, which architecture will still stand in three years, whether a vendor's quote is fair, what the roadmap should be, when to make the first technical hire and who. It is the right shape for a company whose technology choices carry real money and real risk but do not yet fill a full week of a full-time executive.
The defining limit is in the word "direction." A fractional CTO advises; the building happens somewhere else, whether that is your own team, a freelancer, or an agency. The judgment is real and valuable, and on its own it produces documents and decisions rather than a running system. Someone still has to turn the direction into the thing.
What does an agency actually do?
An agency is a team that builds. The deliverable is the artifact itself: the site, the application, the integration, made by several people who each bring a craft to it. Agencies are good at exactly what a single advisor is not, namely capacity and range of hands, and for a defined project that is precisely what you want.
The defining limit here is the other half of the same coin. An agency is usually pointed at a scoped project with an end date, and ongoing direction across the whole of a business's technology is not typically what it sells or what it is structured to own. It will build well what it is asked to build. Whether the right thing was asked for is generally treated as the client's question, not the agency's.
The split that quietly causes the trouble
Set those two limits side by side and an obvious-looking answer appears: hire a fractional CTO for the direction and an agency for the delivery, and let the two combine into a result you own. Sometimes that works. Often it produces the failure this whole piece is about.
The CTO advises. The agency builds to the spec it was handed. Months later the result underperforms, and each party has a true story that points at the other: the specification was wrong, or the build was wrong. Direction sat in one place and delivery sat in another, and accountability for the outcome sat in neither. The owner is left holding a system that nobody will fix without a third invoice, because no single party ever owned whether the thing actually worked.
Direction in one place and delivery in another adds up to a result that nobody owns.
Who actually owns the outcome?
So the useful question is not "individual or team," and it is not even "advice or building." It is this: when the result goes wrong, whose problem is it? Whose name is on the outcome rather than on the hours?
A setup where one party owns both the direction and the delivery has a clean answer to that question. A split setup has either no answer or two competing ones, which in practice is the same as none. This is why the framing at the top is a trap. It sorts the options by what they are called instead of by where accountability lands, and accountability is the only variable that reliably predicts whether you end up with a working system or a standoff.
| Dimension | Fractional CTO | Agency | One party holding both |
|---|---|---|---|
| What it sells | Direction: decisions, not code | Delivery: the artifact itself | Direction and delivery together |
| Owns the outcome | The advice, not the running system | The build to spec, not whether it was the right spec | Whether the thing actually works, end to end |
| Continuity | Inside one head, a single point of failure | Rotates people, context can evaporate between phases | Held as a matter of structure |
| Cost shape | Monthly retainer, no salary or severance | Priced to a defined project scope | The two combined, measured against the cost of a result nobody owns |
| Right when | You already have a delivery team | You already have clear direction and a scope | The outcome is ambiguous and needs both |
Continuity: who remembers why
Whichever way you go, the value compounds when the same party carries the context month after month and remembers why last quarter's decision went the way it did. A fractional CTO gives you that continuity inside one head, with the matching risk that one head is a single point of failure when the person is busy, unwell, or moves on. An agency gives you capacity, but it often rotates people through your project and treats it as one of many, so the reasoning behind earlier choices can quietly evaporate between phases.
The arrangement that holds both, the memory of why and the capacity to act on it, is the one where the same accountable party decides and delivers and keeps the context as a matter of structure rather than goodwill. Continuity stops depending on whether the one person who understood your system happens to still be around.
Advice is not delivery, and delivery is not direction
The two failure modes are mirror images. A recommendation that nobody is resourced to execute simply stalls, a stack of sound advice that never becomes anything. A build with no one owning the direction ships the wrong thing, efficiently. The first is the risk of buying direction alone; the second is the risk of buying delivery alone.
What closes the gap is the ability to put real work behind a decision. When a recommendation needs proving, a load test, a migration trial, a security review, the party that can act on it in days settles the question, while the party that can only advise hands you one more document and a suggestion to find time. The point of holding both halves is that direction gets tested against reality instead of filed.
When one or the other on its own is the right call
None of this means you always need both. Often you need exactly one, and paying for the other is waste. If you already have a capable delivery team and just need senior judgment a few times a month, a fractional CTO on its own is the right and cheaper answer. If you already have clear direction, your own or a trusted advisor's, and you need something built to a defined scope, an agency on its own is right, and it will build it better than a half-resourced generalist ever could.
The split only becomes a problem when the outcome is genuinely ambiguous and you divide ownership of it anyway. The honest test is whether the result has a clear definition of done that someone other than you is accountable for. If it does, a single specialist is fine. If it does not, dividing the work divides the blame and leaves the outcome ownerless.
What does each cost?
Price follows the shape of the help, and knowing the rough ranges keeps the cost from driving the decision by surprise. A fractional CTO is usually bought as a monthly retainer, commonly from a few thousand to around fifteen thousand US dollars a month depending on the seniority and the cadence, with no salary, no search, and no severance attached. An agency is usually bought as a project priced to a defined scope, or as a monthly build retainer, and the figure depends entirely on what is being built. The two are not really substitutes on price, because one is buying direction by the month and the other is buying delivery by the project.
The arrangement that holds both, direction and delivery under one accountable party, is priced as a combination of the two, and the instinct to read that as simply "more expensive" is usually the wrong frame. The comparison that matters is not retainer against retainer. It is the combined cost against the cost of the failure this whole piece is about: a result nobody owns, a standoff between two parties, and a third engagement to repair what the first two could not agree on. Measured against that, paying one party to own the outcome is the cheaper path more often than it looks.
What none of these should ever be is a number with no defined scope behind it. A fractional retainer should say what cadence and which decisions it covers. An agency project should say what it builds and by when. The combined arrangement should say, in writing, who is accountable for the outcome. Price is the easy question. Scope and accountability are the ones that decide whether the money was well spent.
How do you tell which you need?
Two questions cut through most of the confusion. First: when a technology decision goes wrong today, who carries it? If the honest answer is "the owner, alone," the gap is not a shortage of advice. It is that nobody whose actual job is to carry the outcome is in the room. Second: does the work in front of you need someone who decides, someone who builds, or both at once?
If the answer to the second question is "both," the cleanest structure is one accountable party that holds direction and delivery together, so the two cannot drift apart and the result has a single owner from the first call to the running system. If the answer is "just one," buy just one, deliberately, and do not manufacture a split you will have to referee later.
When the decision in front of you is larger than "who should build this" and closer to "should we build this at all," the cleanest first move is an independent read from someone with no stake in the answer. The engagement below is exactly that: scoped advisory, fixed in time and price, with a written recommendation an owner can take to a board and defend.
