Delivery process
This is the path a piece of work takes from a first conversation to a clean exit. The shape is the same whatever the engagement: understand the ground, agree what good looks like, deliver in the open, and leave the client stronger than we found them. What changes is the depth of each stage, not the order. The binding version of these stages, with the controls at each one, is the Standard Operations Policy and Procedure; the practical checklist for a person starting on a client is Joining an engagement.
Discovery and shaping
Section titled “Discovery and shaping”Before we commit to a plan, we get our bearings. That means understanding the landscape as it actually is: the data and platforms in play, how they are built, how they are governed, and who relies on them. From there we set out our observations and the high-level recommendations that follow, and then we shape the work: what we would do, in what order, and why.
Estimates come out of this stage, not before it. We size work from what discovery told us and from experience of similar engagements, and we are honest about what we cannot yet size. A phased shape, where an early phase de-risks the later ones, is usually better than one big number.
Kick-off
Section titled “Kick-off”Kick-off is where an agreed plan becomes a running team. A few things have to exist before it is worth starting:
- A signed statement of work, with scope, outcomes and the commercial basis agreed.
- Access sorted: the accounts, environments and systems the team needs, and a named client contact who can unblock the rest.
- People confirmed on both sides, with roles and availability clear.
- A capability baseline, so we know where the client is starting from and can measure the uplift later.
Kick-off itself stands the team up: we map the key people and dependencies, agree how we will work together, and set up the boards, environments and rituals the delivery runs on. If a prerequisite is missing, that is a reason to hold kick-off, not to push through it.
We also speak the client’s language from the start: their names for their meetings, roles and stages, not ours. If one of our terms would genuinely help, we offer it later, carefully, and only where it adds something.
Running the work
Section titled “Running the work”We deliver in the open. Work runs on an agile basis, usually a kanban flow, against a visible plan on a page and a single backlog the client can see. A RAID log tracks risks, assumptions, issues and dependencies. Write each assumption and dependency with the risk it carries (“the beta slips if the client’s team shrinks” rather than “team size stays constant”), so the “so what” is never left implicit. A burn-up or equivalent shows progress against the goal rather than just activity.
Reporting has a steady rhythm and a plain shape: progress, challenges, wins and asks. Steering starts frequent, often fortnightly, and relaxes to monthly as trust builds and the work becomes predictable. When something needs a decision above the team, it goes to steering with a recommendation attached, not just a problem. Scope is re-planned openly when priorities change, rather than quietly absorbed.
The delivery lead owns the cadence and the reporting. The client sponsor owns the decisions that are theirs to make. Keeping those two clear is most of what keeps an engagement healthy.
Two habits keep that relationship honest. We put ideas to the client as intentions rather than done deals, so there is room to be told no. And when we think the client is about to make a significant mistake, we say so in terms of business risk or cost; if they decide otherwise, we record the decision plainly and factually, and get behind it.
Handover
Section titled “Handover”We plan the exit from the start, because the point of the work is a client who no longer needs us for it. Handover is not a document dump on the last day. Through delivery we transfer ownership progressively and uplift the client’s own capability, so that by the end the team can run what we built.
At the close we give a structured handover to the people who will own the work, cover any training the end users need, and measure success against what we set out to achieve: the numbers reconcile, the thing is adopted, and the people using it say it is better. Then we get out of the way.