A scoped engineering advisory engagement is a closed, written contract for a specific piece of senior technical judgment — one decision, one question, or one artefact — delivered inside a named window and then ended. The deliverable is agreed before the work starts. The end date is agreed before the work starts. There is no standing monthly invoice, and no ambient obligation that survives the last memo. Most operators who think they want a retainer with a senior engineer in fact want this shape, and discover the fact after the first one.
The word retainer is the source of the confusion. In its original legal sense a retainer was an undertaking — the senior practitioner agreed to be available, was paid for the availability, and billed against it when called on. In its modern technology-services sense, the word has drifted into something much vaguer: a monthly standing fee against which the studio does, broadly, whatever the client asks, for as many hours as are reasonable, with the scope rewritten by email as the quarter unfolds. The price is steady. The deliverable is not. The end is not named. The arrangement renews by default, and terminating it has the social weight of a break-up rather than the clerical weight of closing a file.
We have watched good operators take a retainer in that drifted modern sense and regret it eighteen months later. The regret is rarely about the senior engineer on the other end — that person is usually still doing serious work. The regret is about the shape: about the fact that the engagement has no natural conclusion, has no written record of what was decided, and has become a line item on the books whose original reason is no longer remembered. The senior judgment is still being delivered. It is simply not being delivered in a form anyone can keep.
What a scoped engagement looks like instead
A scoped engagement has four visible parts, and they are all written down before the first invoice is issued. There is a question — the specific decision, the specific read, or the specific artefact the work will produce. There is a window — the number of weeks, calendar-bound, inside which the work happens. There is a deliverable — a memo, a decision record, an architectural review document, or a named set of the same. And there is an end — the date on which the engagement is closed and the final payment is made, after which neither side has a standing claim on the other.
The writing-down of these four parts is the work. It is common, in the first conversation, for the operator to describe the engagement in terms of hours or months — a few hours a week of the studio's time, for a quarter, to be on hand. Our reply is almost always to ask what decisions are on the table for that quarter and which of them would benefit from a written senior read. The conversation then reshapes itself around two or three specific reads with named deliverables, rather than an ambient availability. The total cost is often similar. The total clarity is not.
One consequence is that the engagement can be priced by its shape rather than by its duration. A scoped read on a build-versus-buy decision is a different engagement from a scoped read on a post-incident architectural review, and both are different from a scoped diligence for a family office considering a follow-on investment. The shape is legible. The letter is short. The operator can compare the engagement to what a law firm would do for a single opinion, or what an accountant would do for a single filing, and recognise the grammar.
A retainer pays for availability. A scoped engagement pays for a named piece of judgment. Most operators discover they wanted the second only after paying for the first.
The other consequence is that the relationship can continue without becoming ambient. A client who has engaged the studio for three scoped reads over a year has a written record of three pieces of senior judgment, the reasoning behind each, and the point at which each engagement was closed. The next engagement begins with a new question and a new letter. The client ends up owning a small library of decisions rather than a relationship with a perpetually renewing subscription. When a new senior engineer joins the client's team, they read the memos, not their predecessors' chat archives.
Where the open-ended retainer is still right
We should be fair to the other shape. There are engagements for which an open-ended arrangement is the correct structure, and three in particular are worth naming so the distinction is clear.
The first is the embedded fractional executive — a fractional CTO or head of engineering taking a standing role in the company's leadership, attending the same meetings each week, making the recurring calls the business needs made. That is not counsel work. It is a legitimate and separate category of engagement, and we have written elsewhere about why we do not confuse it with what the studio does.
The second is a genuine on-call relationship in which the client needs to be able to pick up the phone on twelve hours' notice and talk to the same senior engineer who has read their system before. A small number of clients do need that, and for them a documented availability retainer — with a written scope, a named senior reader, and a bounded monthly fee — is the right shape. It is rarer than operators believe; most operators who think they need it benefit more from two scheduled scoped reads a quarter and a clear rule for when an out-of-cycle call is warranted.
The third is an ongoing coaching engagement with an in-house team, which is a recognisable shape in its own right. It has a clock, an end, and an explicit transfer of judgment from the external senior to the in-house team; it is a retainer in the older sense rather than in the drifted modern one, and the studio is happy to write that kind of letter when it is the right fit.
Outside those three, the ambient-retainer shape tends to survive because it is easier to sell than to justify. The studio keeps a standing invoice; the client keeps a name on the roster; neither side is forced to defend the continuing value of the arrangement in plain terms. That suits the provider rather more than it suits the operator.
The studio's stance
We prefer scoped engagements and we will propose them by default. When an operator asks us for a retainer, we reply by asking what the next two or three decisions are, and we offer scoped reads against each. When an operator insists on an ambient retainer and no specific decisions are on the table, we decline politely and recommend a firm whose shape fits what the operator has described. We are aware that this loses us revenue we could, in principle, invoice for. We are also aware that in three years' time the operator would not be able to defend, to a new board member or to a future buyer, what the monthly line item had been for — and that an engagement whose value cannot be defended in plain English three years later is an engagement we would rather not have sold in the first place.
The retainer that is not a retainer, in our practice, is the scoped engineering advisory engagement. It is closed. It is written. It ends. When the next question comes, the operator writes to the studio, and, if the shape is right, we begin again under a fresh letter, from the start.