What the first week of an engagement actually looks like

The week is a design, not a warmup

Most buyers have been through a kickoff that produced a slide deck, a Slack channel, and a sense that “work has started.” A week later, the only new artefact is a list of questions that should have been asked before the contract. That is an expensive habit.

We treat the first week as a designed slice of the engagement. By Friday, both sides should be able to point at a written outcome, a named first increment, the access that exists, the access that does not, and at least one risk that would change sequence. If those are missing, we have not started. We have been busy.

Days one and two: make the work visible

The first two days are about seeing the system as it is, not as the proposal described it. We request the smallest set of access that lets us do that. We write the outcome in a sentence the business already uses. We list the people who will get the call when the workflow fails.

• A single written outcome. Not a backlog. The change that would make the engagement worth doing.
• A map of what we can already see: environments, data stores, third parties, the spreadsheet that is still the source of truth.
• A named operator. If nobody owns the path after we leave, that is a first-week finding, not a last-week surprise.
• A list of access still missing, with who can grant it and what is blocked until they do.

Days three to five: the thinnest useful slice

The rest of the week is spent choosing the first increment that can be finished and reversed. That might be a measured diagnosis, a dual-write of one field, a single integration with a quarantine, or a product path that can be operated by someone other than the founder. It is rarely “set up the whole platform.”

We write down what we will not start yet. The items that feel urgent and are not on the critical path go on that list in public. Scope that cannot be named cannot be protected.

A first week that produces a smaller, truer plan is more valuable than a first week that produces activity.

What you should have in your hands on Friday

• The outcome, in writing, in language the business already uses.
• The first increment, with a definition of done that includes failure, not only the happy path.
• The risks that would change sequence or price, including the ones we cannot see yet.
• How we will talk during the rest of the work: who answers, how fast, and where decisions are recorded.

If the first week reveals that we are the wrong team, that is a successful week. It is cheaper than discovering it in month three. Complimentary conversations exist so that this is rare. The first paid week exists so that it is still possible.

What we will not do in week one

• Rewrite the system. Week one is for learning which part is allowed to change.
• Introduce a new architecture because it is current. We inherit the operating environment you already have.
• Disappear into a private backlog. If you cannot see the next increment, we are not done with the week.

If you are deciding whether to start, ask any partner what Friday of week one looks like. If the answer is only “we will be ramping up,” you are buying a calendar, not a plan.