Working Together

Month One With a Development Agency: What Should Actually Happen

7 min read | Business

You have signed. The kickoff call is booked. Four weeks from now you will either have a product taking shape or a folder of documents and a vague feeling that not much happened.

Month one sets the entire trajectory of a build, and it is the period clients have the least visibility into. Here is what should actually be happening, week by week, and the specific artefact you should be able to point at each Friday. If you have not sent your brief yet, start with how to brief an app development agency — everything below assumes that document exists.

1. Week One — Alignment and Access

Week one is not “getting to know each other.” It is a controlled information transfer, and it should feel slightly intense.

Day 1–2
Kickoff, stakeholder map, decision rights confirmed in writing
Half day
Day 2–3
Access: repos, analytics, staging, design files, third-party accounts
Ongoing
Day 3–4
Technical audit of what exists — or architecture options if greenfield
1–2 days
Day 4–5
User research plan, or synthesis of research you already have
1 day
Friday
Written assumptions log — everything the team is taking on faith
Deliverable
The week-one artefact

An assumptions log. A plain list of everything the team currently believes to be true and has not verified. If your agency cannot produce one by Friday of week one, they are not thinking about risk — and every unverified assumption is a candidate for a change request in week six.

The most common week-one failure is on your side, not theirs: access takes eight days instead of two. Chase credentials before kickoff, not after. Idle senior engineers still bill.

2. Week Two — Narrowing

Week two is where a good team starts telling you no. The feature list gets cut, the architecture gets a decision attached, and the first real estimate replaces the sales estimate.

🎯
Scope triage
Every feature sorted into launch / fast-follow / never. “Never” should have at least a few items in it, and you should be mildly annoyed by one of them.
🏗️
Architecture decision record
One page per major choice: what was picked, what was rejected, and why. Six months later this is the document that stops you relitigating.
📐
Core flows wireframed
Grey boxes, not visual design. Two or three critical journeys, clickable enough to argue about.
📊
Revised estimate
The number now based on what the team has actually seen. If it has not changed at all from the proposal, discovery did not happen.

3. Week Three — First Working Software

This is the week that separates agencies. By the end of week three there should be something running that you can open in a browser or on a device — deployed, not on someone’s laptop.

It will be ugly and it will do almost nothing. That is correct. The point is that the pipeline is real: a commit goes to a repo, CI runs, a build lands in a staging environment you can reach. Teams that defer this until “when there’s something to show” routinely discover deployment problems in the final week, which is the most expensive possible time to find them.

On track by end of week 3
Staging URLLive, you have access
CI/CDRunning on every push
DesignDirection approved
ScopeWritten, signed, cut
DemosWeekly, same slot
Drifting
Staging URL“Soon”
CI/CDNot set up yet
DesignThird moodboard round
ScopeStill “flexible”
DemosRescheduled twice

4. Week Four — Rhythm and the Honest Conversation

By week four the project should have a heartbeat: a fixed demo slot, a visible board, a change-request process that has been used at least once. The delivery cadence matters more than the delivery speed — predictable teams let you plan around them.

Week four is also when a good agency has the uncomfortable conversation. They have now seen your codebase, your stakeholders and your constraints, and they know something the proposal did not. Maybe the timeline needs another three weeks. Maybe a feature you love is technically disproportionate. Hearing it in week four is a gift; hearing it in week eleven is a crisis.

🚩
No bad news yet
Four weeks into any real project, something has gone sideways. Silence means they are managing your perception, not the risk.
🚩
Demos are slides
Screenshots and Figma frames are not progress. You should be clicking something.
🚩
You cannot see the repo
You are paying for the code. Read access from day two, no exceptions, whatever the tooling.
🚩
The team changed quietly
The senior engineer from the pitch is now “overseeing.” Ask who is committing, and check.

5. The Month-One Scorecard

Print this. At the end of week four you should be able to tick every line.

A written, signed scope with things deliberately excluded
Including at least one feature you argued about and lost.
A staging environment you can open right now
On your phone, without asking anyone.
Repository access and a visible commit history
With the names you were promised in the pitch.
Architecture decisions written down with rejected alternatives
One page each. Nobody remembers this in month five otherwise.
A revised estimate based on what they have seen
Up or down — but different from the proposal.
A change-request process that has been used once
Untested processes fail exactly when you need them.

If you are still weighing whether to run this in-house instead, our honest comparison of in-house teams and agencies covers the trade-offs — and if you are arriving with a prototype that is nearly there, the MVP gap describes the month-one that situation actually needs.


Month one should produce software.

Tell us what you are building and we will show you exactly what your first four weeks would look like — deliverable by deliverable, with the uncomfortable conversation scheduled up front.

Start the Conversation →

Engineering Insights

Latest from Syntaxa Studio.

Loading latest posts