TL;DR: The first 90 days with a new development agency determine whether you'll be satisfied or frustrated for the rest of the engagement. Week 1-2 should focus on discovery and team integration. Weeks 3-6 produce the first deliverables and establish working rhythm. Weeks 7-12 hit full velocity and deliver the first major milestone. Here's a realistic week-by-week timeline with the pitfalls most clients miss.
The transition gap nobody prepares for
You've spent months selecting an agency. Contracts are signed. Both sides are excited. Then... silence. Or worse, a flood of meetings with no visible output.
The first 90 days make or break the relationship because expectations are at their highest and context is at its lowest. The agency doesn't know your business. You don't know their process. If this gap isn't closed deliberately, frustration builds fast.
Here's what should happen — and what to watch for.
Weeks 1-2: Discovery and setup
What happens
The agency runs a structured discovery phase. This isn't the team "getting ready" — it's the most important part of the entire project.
During discovery, the agency should:
- Interview your key stakeholders (product owner, technical lead, end users)
- Map existing systems and integrations
- Review any existing code, data structures, or APIs
- Define the project scope with prioritized requirements
- Produce a technical architecture proposal
- Set up development infrastructure (repositories, CI/CD, staging environments)
You should provide:
- Access to relevant stakeholders for interviews
- System documentation and credentials for existing tools
- A named product owner who can make decisions
What to watch for
Red flag: The agency skips discovery and starts coding in week one. This means they'll build based on assumptions instead of understanding. Rework is coming.
Red flag: Discovery produces a 60-page document nobody reads. Good discovery outputs fit on 5-10 pages with clear diagrams, defined scope, and explicit assumptions.
Green flag: The agency pushes back on some of your requirements during discovery. "We've built similar features before, and here's why this approach will cause problems in six months" — that's the kind of input you're paying for.
Deliverables by end of week 2
- Project scope document with prioritized features (must-have / nice-to-have)
- Technical architecture overview
- Sprint plan for the first 6 weeks
- Communication plan (tools, meeting cadence, escalation paths)
- Development environment set up and accessible to your team
Weeks 3-6: First sprints and working rhythm
What happens
The team starts building. The first sprint or two produce smaller, foundational deliverables. Don't expect a fully working product yet — expect to see progress you can review.
Typical first-sprint outputs:
- Core infrastructure and deployment pipeline
- Authentication and basic user roles
- Database schema and API scaffolding
- First UI screens (functional, not polished)
The working rhythm should settle into:
- Sprint planning at the start of each 2-week cycle
- Daily standups (15 minutes, async or synchronous)
- Mid-sprint check-in if needed
- Sprint demo and retrospective at the end of each cycle
Your role as a client
This is where your product owner earns their keep. The agency needs:
- Decisions within 24-48 hours. Not 2 weeks. Decisions bottleneck everything. When a developer needs clarification on a feature and can't get an answer for 10 days, you've just wasted 10 days of their sprint.
- Review sprint deliverables. Actually click through the demo. Provide specific feedback: "This form should validate email format" is useful. "Make it more intuitive" is not.
- Prioritize backlog continuously. New requirements emerge. Bugs get discovered. Features need reprioritization. That's normal. Your job is to manage priorities, not scope everything upfront.
What to watch for
Red flag: The agency doesn't demo at the end of each sprint. If they can't show working software every two weeks, something is wrong with their process.
Red flag: You're not hearing from the team between sprint demos. Good agencies proactively communicate, especially when they discover blockers or risks.
Green flag: Velocity increases from sprint 1 to sprint 2. The team is learning your codebase and domain faster than they initially expected.
Green flag: The agency suggests cutting a feature because it's not worth the complexity. They're building your product, not padding hours.
Deliverables by end of week 6
- Working prototype or first functional increment (deployed to staging)
- Revised timeline and estimates based on actual velocity
- Identified risks or scope adjustments
- Documented technical decisions and trade-offs
Weeks 7-12: Full velocity and first milestone
What happens
By week 7, the team should be at peak velocity. The discovery learnings are applied, the codebase is established, and the working rhythm is steady.
Expect these patterns:
- Sprint velocity stabilizes (you can predict how much gets done per cycle)
- Fewer clarification questions from the agency (they understand the domain)
- More proactive suggestions from the team about features and improvements
- First deployable milestone — an MVP, internal beta, or proof-of-concept
The milestone review
Around week 10-12, schedule a formal milestone review. This is not a sprint demo — it's a business-level assessment:
- Is the project tracking against the original goals and budget?
- Are there scope changes that need re-estimation?
- Does the team composition need adjustment (more frontend, less backend)?
- What's the realistic timeline for the next milestone?
This review should feel collaborative, not adversarial. Both sides share data, discuss what worked and what didn't, and align on the path forward.
What to watch for
Red flag: The agency hasn't deployed anything you can test. After 12 weeks, you should have a staging environment with working software. Not finished software, but working software.
Red flag: The team composition changed without discussion. If the developer you met during the sales process isn't on your project anymore, ask why.
Green flag: The agency identifies technical debt and proposes a plan to address it. They're building for the long term, not just sprinting toward a demo.
Communication: the make-or-break factor
Recommended tools and cadence
| Activity | Frequency | Tool suggestion |
|---|---|---|
| Daily standup | Every workday, 15 min | Slack, Teams, or async Loom video |
| Sprint planning | Every 2 weeks, 60-90 min | Zoom/Meet + Jira/Linear |
| Sprint demo | Every 2 weeks, 30-45 min | Zoom/Meet with screen share |
| Retrospective | Every 2 weeks, 30 min | Zoom/Meet |
| Escalation | As needed | Direct call to project manager |
| Progress reporting | Weekly | Email or Slack summary |
Async vs. synchronous
If your team and the agency are in different time zones, lean toward async communication:
- Written sprint summaries instead of long meetings
- Loom videos for walkthroughs and demos
- Slack for quick decisions (with a 4-hour response time expectation)
- Synchronous calls for planning, major decisions, and demos only
The agency should adapt to your communication preferences, not the other way around. If you hate meetings, say so. If you want a daily update email, request it.
The 90-day health check
At the end of 90 days, evaluate the engagement against these criteria:
| Criteria | Healthy | Warning |
|---|---|---|
| Deployable code on staging? | Yes, multiple sprint increments | No working deployment |
| Sprint velocity trend | Stable or increasing | Declining or erratic |
| Decision turnaround time | Under 48 hours | Over one week |
| Team stability | Same core team since week 2 | Multiple team changes |
| Communication quality | Proactive, transparent | Reactive, vague |
| Scope managed? | Changes discussed and re-estimated | Scope creep with no visibility |
If three or more items fall in the "Warning" column, it's time for a direct conversation about what needs to change.
Frequently asked questions
How involved do I need to be as a client? Plan for 5-10 hours per week from your product owner during the first 90 days. Sprint planning, demos, backlog grooming, and ad hoc decisions. This drops to 3-5 hours after the team is established. If you can't commit a product owner, the project will drift.
What if the agency's first proposal after discovery looks different from what I expected? Good. It means discovery worked. The agency now understands your problem better than you described it in the RFP. Evaluate the proposal on its merits. If the changes are justified and explained clearly, that's a sign of a strong agency.
How do I handle disagreements with the agency during the first 90 days? Raise them early and directly. The first 90 days is when you establish the working relationship. If you swallow frustrations, they compound. A good agency welcomes direct feedback — it helps them calibrate to your expectations.
What does Globalbit's onboarding look like? We start every engagement with a structured discovery phase (2-4 weeks). Your designated product owner gets a named project manager and tech lead from day one. Daily standups start in week one, and you'll see deployed code on staging by the end of sprint one. After 150+ projects, we've refined this process to eliminate the common ramp-up failures. Let's walk you through it.