Technology Leadership As A Service, Explained

A harbour pilot boards a ship that is not his own, takes the con for one stretch of water, and is off the bridge before the open sea. The master keeps the ship, the crew and the voyage; the pilot brings local knowledge and, for that stretch, the authority to use it. Technology leadership as a service runs on the same bargain, and it comes apart in the same place: nobody sets down in advance where the stretch begins and where it ends.

Four arrangements travel under this one label. One is a board adviser who joins a monthly review and offers an opinion on request. Another is a part-time chief technology officer who works inside the engineering team for a set number of days each month. A third is an appointed executive who occupies a vacant seat for a fixed period. A fourth is a specific leader with a delivery organisation behind them. Until the client knows which of the four is on the table, two offers cannot be compared.

Decision capacity is countable, and the count is the unit of comparison

Every engagement of this kind is there to move decisions that nobody inside the business is placed to move alone. Hiring, platform choice, build against acquire, security posture, whether a legacy system lasts the next budget cycle: the set is short and specific, and it differs in every company. Set it out before you speak to anyone. A count of decisions per quarter, each with a deadline beside it, is the only unit that lets two offers be compared honestly, because a day measures the provider's effort and a decision measures your exposure. DJH Tech Partners sets out its board-level advisory as translating technical complexity into business decisions for a senior team, and the translation is the easier half. The harder half is agreeing which decisions belong inside the engagement, because an engagement that never identifies them will default to advice.

The four terms that turn a brief into a delegation

The mandate is one or two pages, signed before work starts, and it is a delegation of authority. Four terms do most of the work.

The decision term sets out which decisions the leader takes alone, which come back to you, and which belong to somebody else on the team.
The substitution term sets out who does the work when the appointed individual is unavailable, and how much notice you get of a change of face.
The exit term sets out what is handed over on the last day, in what format, and who confirms that it is usable.
The cadence term sets out the standing review, the three numbers read at it, and the brief issued after each cycle.

Everything else, the fee included, becomes easier to agree once those four terms are in place. A provider who resists setting them out is telling you how the engagement will be run.

Four arrangements placed against the attributes that decide fit

The same four attributes settle fit in every serious proposal, and the label across the top of the offer tells you least of all. Read the table down a column.

Arrangement	What is acquired	Who owns the decision	What the clock measures	What remains after the final bill
Advisory	An expert view on stated questions	Your own executive, advised	The review cycle	A record of the recommendation and the reasoning behind it
Embedded fractional	Presence inside the technology function	Shared, under a signed delegation	Days per month and the standing review	The team, the architecture record, the prioritised backlog
Interim	The vacant seat itself	The interim, as the appointed executive	The notice period and the search for a successor	The handover file and the person who replaces them
Programme or managed	A specific leader with a delivery organisation behind them	That leader, inside an agreed scope	Milestones and the work package	The delivered system, its runbook and its support model

Your half of the arrangement

No provider can supply the three things that decide whether the engagement produces anything. The first is access: the leader needs to be in the room with the people who own the constraints, which means finance, sales, security and the engineering leads, and a calendar that lets those conversations happen inside the paid time. The second is one internal owner, a specific person on your side who takes each decision into the business after the review ends. The third is a decision log that anyone on the team can read, so the reasoning outlives the engagement and whoever takes the seat next does not start from nothing. Arrange all three before the first payment leaves your account, and the arrangement you chose matters far less than the discipline you brought to it.

Why hourly arithmetic hides the difference between two offers

Two offers that look identical on hours are selling different things. NorthBay Solutions provides a specific virtual leader for 20, 30 or 40 hours a week, for 30 or 60 days at a time, backed by a bench of engineers certified on AWS, which suits a defined project with a closing point. Rational Partners runs a different model: experienced operators working fractionally, drawn from a bench of more than 25 CTOs and CPOs across the UK and Europe, with fewer than one in five candidates selected. One is a capacity engagement and the other is an advice engagement, and the hourly arithmetic that makes them look comparable is exactly what hides the difference. Count the decisions in the mandate, divide by the cadence at which they have to be taken, and the arrangement usually selects itself.

Pricing the mandate against the cost of delay

A full-time technology leader costs a salary, employer contributions, a search fee and, most expensively, the risk of choosing badly. The day rate decides nothing on its own. What decides is the cost of the decisions in the mandate, set against the cost of those decisions arriving late or wrong. Where the mandate covers four hiring decisions and one platform decision that between them settle the next year of revenue, the fee is a rounding error and the mandate is the whole conversation. The technology code of practice, published by the Government Digital Service and the Central Digital and Data Office, is a set of criteria for designing, building and procuring technology, and it is a useful public yardstick when you want to know whether a proposal would withstand scrutiny from outside your own company.

What the supply side is actually selling

Providers divide by what stands behind the individual who does the work, and the difference shows up in the first month. Rational Partners sells a firm, with pattern recognition built across more than a hundred engagements and colleagues the fractional CTO can consult when a problem is unfamiliar. NorthBay sells delivery capacity. Etteplan says its industrial service suits work needing at least a year of effort, and that the role can look like a development director in a large business and a temporary CTO or CDO in a smaller one. Smartdesc's IT leadership service takes accountability for an IT function, with virtual IT directors who report at executive level and work to an IT health scorecard refined over ten years, and its engagements often begin with more days and scale back once the function is steady. INNOSEC draws the line that matters here: management keeps the lights on by fixing laptops, configuring Microsoft 365 and responding to incidents, while leadership sets direction and ties systems to what the firm is trying to achieve. The title at the top of the contract tells you least of all, so ask what stands behind the person who signs it.

Demand for the model is being pushed from outside the market as well. Bristow Holland's report on the changing technology leadership role sets out the Department for Business and Trade's SME Digital Adoption Taskforce recommendation of a scalable CTO-as-a-service model, with AI-powered guidance for smaller firms. That attention signals demand. It does not settle the terms of an engagement, so the mandate still decides.

Checks at thirty days, ninety days and at the end

The arrangement is working if three things are in place by day thirty: entries in the decision log, one person on your side who owns each decision after the review ends, and a calendar that places the leader in front of the people who hold the constraints. If the log is empty at thirty days, the engagement has become a series of conversations, and the fee is paying for conversation.

By ninety days, four questions settle whether it continues. Is the decision list shorter, because the calls have been made, or longer, because the work kept uncovering more of them? Is the internal owner still the same person? Could another engineer read the log and follow the reasoning without asking anyone? And would the next quarter look materially different if the engagement stopped on Friday?

At the end, the test is the handover. The pilot leaves the bridge at the seaward mark, and the master takes the ship into open water with everything the pilot knew already aboard. Ask for the same: the architecture decisions with their reasoning, the hiring plan with its sequence, the risks that were accepted and who accepted them, and the short set of things that will break if nobody watches them. A provider who cannot produce that material in the final week has been selling presence, and presence is the one thing the fee was never meant to buy.

The stretch of water is the whole point. Mark it on the chart before anyone steps aboard, and the arrangement does the work it was engaged to do.