The three arrangements get compared on hours, as though they were the same thing at different intensities. They are not. They answer different problems, they run on different rhythms, and they end in different places.
Hours are the easiest thing to compare and the least useful. A company can buy a great many hours of the wrong arrangement and end up exactly where it started, having spent the budget it needed for the right one.
The difference is accountability
The useful question is not how much of someone you get. It is who is answerable when a technology decision turns out badly.
In advisory, you are. You bring questions, you get experienced judgment, and you remain the person who decides and owns the outcome. That is the right arrangement when the judgment is what is missing, not the ownership.
In a CTO retainer, I am, alongside you. I hold the technology direction with the founder, which means being in the room when the decision is made rather than consulted about it afterwards, and being wrong in public when I am wrong.
In venture building, I am, across several companies at once. The unit of accountability is not one product but the standard every venture is built to — what they must meet before leaving the studio, and what they may decide for themselves.
Everything else follows from that difference.
What each one is actually for
Advisory fits a company with capable technology leadership that lacks a second opinion. A founder who is technical and busy. A lead engineer who is strong but has not built at the next scale. A board that wants a view it did not commission from the team being assessed. The work is pressure-testing: an architecture decision while it is still cheap to change, a vendor's AI claim before it is bought, a readout before an investor conversation.
It is also the right first arrangement when nobody is sure which arrangement is needed, because it costs least to be wrong about.
A CTO retainer fits a company being built AI-native that has no senior technology leadership above the team. The work is continuous rather than episodic: the architecture and data decisions agents depend on, where people stay accountable and where they do not, evaluation and guardrails, and the ordinary unglamorous business of making a roadmap survive contact with reality. It is the arrangement where someone else's judgment becomes part of how the company runs, rather than something it consults.
Venture building fits a studio rather than a company. The problem is not that one venture lacks leadership; it is that every venture re-decides the same questions and arrives at different answers, so the portfolio's technical quality is whatever each founding team happened to know. The work is the standard, the governance of AI-assisted engineering across teams, and the first ninety days of each new venture's technology.
How each one feels
Advisory feels like access. You have someone to ask, the answers are quick, and the responsibility stays where it was. If you find yourself wishing the adviser would simply go and fix it, you have outgrown the arrangement.
A retainer feels like a colleague with a diary. There is a rhythm — planning, decisions, reviews — and you stop preparing for the meetings because the person is already across it. If you find yourself briefing them from scratch each time, the cadence is too thin for the problem.
Venture building feels like infrastructure. Much of the value is invisible from inside any single venture, because it shows up as problems that did not happen: the venture that passed diligence first time, the platform decision that was made once rather than five times.
How each one ends
Advisory ends when the questions stop arriving, which is usually a good sign. It can also quietly continue for years at a low level, which is fine if it is deliberate and a waste if it is inertia.
A retainer should end in a handover. The outcome you want is a company that no longer needs the arrangement: a permanent technology leader hired well, or a team that has absorbed the judgment. An engagement with no theory of its own ending is an expensive habit.
Venture building ends when the standard holds without being held. The studio runs its own technical review, new ventures start at the standard rather than working towards it, and the outside help becomes occasional.
The expensive mistakes
Buying advisory for a retainer problem. The common one, and it usually comes from budget. The company has no senior technology voice, delivery is drifting, the board is uneasy, and it buys a few hours a month of good advice. The advice will be good. Nobody will own it, and the situation will keep deteriorating, because advice is not a substitute for accountability.
Buying a retainer for a studio problem. Less common and more wasteful. A studio brings senior leadership into its most troubled venture, which helps that venture and leaves the portfolio exactly as it was. The next venture starts from scratch, because nothing was built that outlives one company.
Buying venture building for one venture. A studio arrangement applied to a single product is an expensive way to get a retainer, with overheads designed for a problem you do not have.
When the answer is none of them
Two situations call for something else.
If the company is in crisis — delivery broken, platform unstable, technology leadership just departed — it needs someone holding the role rather than advising it, with full attention and real authority. That is a different arrangement from all three above, and recognising it early matters more than which of these you would otherwise have chosen.
And if what you need is a decision rather than a relationship — is this AI claim real, should we build or buy this capability, why did the pilot stall — that is a fixed piece of work with a scope and an end date. It is bought like a product, not like capacity, which is why due diligence is priced and leadership is not.
The three arrangements above are all ways of reserving judgment over time. The question is how much of it you need continuously, and who you want answerable when it is wrong.