Delivery Implementation for Engineering Teams

What a six-month engagement is for

The engagement rebuilds how an engineering organisation operates: process, hiring, compliance readiness, observability, on-call, and a platform function that does not exist yet.

Every one of those is a project a competent team could run itself. What almost no team can do is run six of them while also shipping, because each competes with the sprint and loses. Run as separate initiatives they finish at different rates, in different orders, and the ones nobody is measuring quietly stop.

Six months is what the sequence takes when it is somebody’s whole job. The first milestone lands at a month, which is early enough that a client can see whether this is working before most of the money has moved.

Sequenced, not parallel

The order is not arbitrary and it is not negotiable in the way clients often want it to be.

Process comes first because everything downstream is measured through it. There is no way to tell whether a platform team helped if the delivery signal was noise to begin with.

Hiring comes second, because senior hires take two to three quarters to arrive and be useful. Starting them in month four means they land after the engagement ends, which is the same as not starting them.

Compliance runs alongside from month two, because the evidence pipeline has to have been collecting for a while before an auditor finds it credible. This is the track where starting early costs nothing and starting late costs a quarter.

Observability and on-call come next, in that order — a rotation without instrumentation is just a schedule of people being woken up without information.

The platform function comes last, because a platform team chartered before the delivery teams have a shared process will build shared services for practices that are about to change.

Why you can stop

Each milestone has a defined outcome and you are billed on its completion. If it did not deliver, you decline the next one and the engagement ends.

This is not a gesture. A six-month engagement paid up front is a bet the client makes alone, on information they do not have yet, and it produces the failure mode where nobody wants to be the first to say it is not working. Moving the decision to four checkpoints puts it where the evidence is. It also constrains us: every milestone has to be scoped as something that can be shown to have happened, which rules out the tracks that only pay off after the engagement ends.

It ends with a handover

The final milestone is a permanent technical leader taking the seat, usually someone hired through the engagement’s own hiring track.

That is a conflict of interest and it is worth naming: we are being paid to recruit our own replacement. The alternative is an engagement that ends with the client dependent on renewing it, which is the incentive that makes long consulting relationships go bad — and it is the reason the handover is written into the scope rather than left as an intention.

Two engagements sit either side of this one. Fractional CTO is the retainer version: the same seat, held month to month, without the six-track programme underneath it — and the two are often run together, with the retainer holding the leadership seat while this rebuilds what sits below it. Engineering Hiring is the hiring track sold on its own, for teams whose only structural problem is that the loop does not work. Where the programme runs across several teams at once and the risk is coordination rather than capability, that coordination is Technical Program Management.