The three doors.
The assessment runs ten working days. At the end we deliver four artefacts: an architecture decision record, a risk-ranked delivery sequence, a build/buy/partner recommendation with numbers, and a customer-facing technical narrative. Then the customer chooses.
Door one: Build. We take the difficult slice ourselves. Usually this is the integration, the firmware, the AI feature, or the part of the system where the customer's team is missing a specific skill on a fixed schedule. Principal-led delivery, fixed-price where we can scope it, time-and-materials where we cannot.
Door two: Advise. Your team builds. We sit on weekly reviews, unblock specific decisions, and stay close to the technical risks named in the assessment. Capped retainer: a fixed maximum monthly fee, used or not used. The retainer ends when the work ships or when the customer decides the value has stopped compounding.
Door three: Hand back. The four artefacts are yours. You keep the decision record, the sequence, the recommendation, the narrative. You take them to your team, your board, another consultancy, or a contractor pool. We walk. No retainer, no follow-up sales sequence, no "if you change your mind".
Why hand-back is a costly signal.
Most consultancies cannot afford to offer this. The economics of a leveraged consulting firm (partners selling, senior staff scoping, juniors delivering) require follow-on revenue to cover the cost of the assessment. The assessment is a loss leader. The follow-on is the business model. The pressure to convert is built into the pricing.
That pressure produces predictable distortions. Assessments recommend the work the firm is staffed to deliver. Build/buy decisions favour build. Hand-back is not on the menu, because hand-back means the assessment must pay for itself.
The assessment pays for itself here because Thrucible is principal-led. There is no associate margin to feed. The £9,999 is priced as a standalone product. If the customer takes the artefacts and walks, the engagement was profitable. The follow-on revenue is upside, not requirement.
This is the costly signal: we can afford to be wrong about whether we are the right people to build the next thing. That changes how the assessment gets written. It also changes which assessments get sold; the wrong-fit projects screen themselves out.
What the four artefacts buy you.
Each one is designed to keep working after we are gone.
- Architecture decision record. The technical shape of the solution, the alternatives considered, the trade-offs accepted, the constraints assumed. Written so a competent engineer joining the project in week six can understand why the system looks the way it does.
- Risk-ranked delivery sequence. What to build in what order, with named decision points, named technical risks, and the latest reasonable date to escalate each one. Not a Gantt chart: a sequence that names what gets resolved before what.
- Build/buy/partner recommendation with numbers. For each major component, the actual choice with the actual budget. Not "consider exploring vendors". A named recommendation with named alternatives and the comparative cost.
- Customer-facing technical narrative. The version of the plan you can share with your enterprise customer, your board, your investors, or your buying committee. Plain English, no jargon. Designed to close a stakeholder conversation, not start one.
Hand-back customers usually take all four to their internal team. Some take them to another consultancy who delivers against the recommendation. A small number take them to nobody: the artefact set is enough for the customer to decide not to do the project at all. That is sometimes the most valuable possible outcome.
When hand-back is the right choice.
It is the right choice in three situations.
First, when the customer's internal team has the skills but needed an outside read. The assessment confirms the plan they were going to build anyway, sharpens the risk register, and gets them through their board with a credible narrative. They keep the budget, they keep the delivery, they keep the institutional learning.
Second, when our build pricing comes in above what the customer is willing to pay and the work is not the kind we are willing to discount. The assessment names the path. Another delivery partner can execute it. We are not the only competent option, and the customer knowing that is a feature of the relationship.
Third, when the assessment reveals that the project as scoped is not worth doing. The right answer is to stop, refund the assessment fee if the customer wants, and leave them with the artefacts that explain why. We have done this twice. Both customers have come back with different projects.