A well-run agency project moves through discovery, design, development in short cycles, testing, launch and a support period. You should expect a named project manager, a demo of working software every one or two weeks, clear requests for decisions and materials with deadlines, and honest updates when something slips. Your job is to make decisions quickly, provide content and access on time, and show up to demos.
Most founders have never been through a software project before they hire an agency, and the agency has been through hundreds. That imbalance is why it helps to know in advance what a healthy project looks like at each stage, what will be asked of you, and which early signs tell you whether things are on track. This is how ours run, and what any good agency's should look like.
Before kickoff: discovery
What happens: workshops with the project manager and lead developer to define the problem, the users, the first release and the road map. At 7L we call this defining the ultimate problem, and it typically takes one to two weeks. The output is a scoped first release with a fixed price and timeline.
What you provide: your time, your interview notes and validation data, your brief, access to anyone in your business who knows the process being replaced.
Good sign: the team challenges some of your assumptions and proposes cutting scope. Warning sign: they agree with everything.
Weeks 1 to 3: design
What happens: user flows, wireframes and then visual design of the key screens, usually as a clickable prototype you can test with real users. Meanwhile, the developers set up the technical foundations: repositories, cloud environment, the skeleton of the backend.
What you provide: brand assets, fast feedback on designs, and decisions. Design review is the first place projects stall, because founders want to perfect every screen. Aim for "good enough to build" and note improvements for later.
Good sign: designs arrive in small batches with specific questions. Warning sign: a big reveal of every screen at once.
Weeks 3 to 12: development in cycles
What happens: the team builds in short cycles, typically two weeks each. Each cycle has a goal, and at the end of each you see a demo of working software, however rough. Backend, front end and integrations progress together, so early demos show a thin version of the whole journey rather than a polished version of one screen.
What you provide: attendance at demos, feedback within a day or two, content and data the product needs, and access to third-party accounts and any systems being integrated. This is also when change requests arise. Expect a clear process: the request, its effect on price and date, your decision.
Good sign: every demo shows something you could not do last time. Warning sign: demos are postponed, or you are shown slides instead of software.
Weeks 10 to 14: testing and hardening
What happens: a dedicated quality phase overlapping the last development cycles. Automated tests run continuously, testers work through every journey on real devices, performance is checked under realistic load, and security is reviewed. Bugs are found, prioritised and fixed. The app store listings are prepared and a first build is submitted for review.
What you provide: testing by you and a small group of friendly users, the app store account details and listing content, decisions on which minor issues can wait until after launch.
Good sign: a visible bug list, shrinking week by week. Warning sign: "no bugs found".
Launch week
What happens: final production deployment, store approval, monitoring switched on, a launch checklist worked through. The team is on standby for the first days because real users find things test users did not.
What you provide: the go decision, your launch communications, and patience. A quiet launch to a small group before the public one is usually wise.
After launch: support and the next release
What happens: a defined support period for fixes and adjustments, then a move to a retainer or a scoped second release. The first weeks of real usage produce the most valuable feedback you will ever get. A good agency helps you read it and plan what comes next.
What you provide: the metrics you set out to measure, user feedback, and a decision about the next phase.
What the agency needs from you, in one table
| Phase | Your inputs | Your decisions |
|---|---|---|
| Discovery | Brief, validation notes, access to your business | Scope of the first release |
| Design | Brand assets, reference products, feedback within days | Approve flows and key screens |
| Development | Content, data, third-party account access, demo attendance | Change requests, priorities |
| Testing | Your own testing, store account, listing content | What ships and what waits |
| Launch | Communications, early users | Go or no-go |
| After launch | Metrics, user feedback | Next release scope |
How do you know early if the project is on track?
- By the end of week two you have met the whole team and seen a plan with dates.
- By the end of week four you have clicked through a prototype.
- By the end of week six you have seen working software, however thin.
- Every demo since has shown progress you can name.
- When something slipped, you heard about it from the project manager before you had to ask.
If two or more of these are missing, raise it now. Problems at week six are recoverable. Problems discovered at week twelve are expensive.
Your part of the deal
The projects that finish on time have a founder who makes decisions within days, supplies content on schedule, attends every demo and keeps the list of "later" features firmly later. The agency can do everything else. It cannot do that.
Curious how this looks at 7L specifically? Our post on defining the ultimate problem describes discovery, and how long an MVP takes puts numbers on the phases. If you are about to start a project and want to talk through what to expect, get in touch.
Frequently asked questions
How much of my time will the project take?
Plan for two to four hours a week during design and development, more during discovery and launch week. The time is mostly demos, feedback and decisions. Founders who cannot give this should nominate someone who can.
What if I cannot attend a demo?
Ask for a recording and send feedback within a day or two. Missing one is fine. Missing several is how a project drifts away from what you wanted.
What should I do if the project slips?
Ask for the cause, the new date and what changed. Slips from your late decisions or content are on you. Slips from underestimation are the agency's to absorb on a fixed price. Either way, an honest conversation early is better than a surprise at the end.
Can I bring in my own designer or advisor?
Yes, and a good agency welcomes it. Agree at the start who decides when opinions differ, so the team is not caught between two clients.
What happens to the team after launch?
It depends on the arrangement. A retainer keeps some of the same people on your product. A new scoped release may rebuild the team. Ask before launch so there is no gap when the first feedback arrives.