I've sat on both sides of this table: as the agency pitching a client, and as the consultant brought in to clean up after a different agency's failed project. The patterns are remarkably consistent. Clients who get burned almost always missed the same handful of warning signs at the proposal stage.
This is the framework I'd give a friend about to choose a web development agency for a $20k–$100k build. It assumes you've narrowed it to two or three. The goal is to find out which one (if any) is right before you sign.
What a failed project really costs
Most clients only count the first number. The third cost is opportunity: every month your site doesn't ship is a month it doesn't bring in customers. For a business doing real lead generation, that's often the biggest number of all.
The five hard red flags
If you see any of these, don't sign. The rest of the pitch doesn't matter.
- Q1Did they quote a number after a 30-minute intro call, with no real discovery?YesWalk away. They're guessing, and they'll cut corners or bill scope changes later.
- Q2Are their recent client sites slow or broken on your phone and in PageSpeed Insights?YesWalk away. If they can't ship a fast site for their clients, they won't for you.
- Q3Are they vague about exactly who will do the work?YesWalk away. A senior pitch with a junior build is a different product.
- Q4Does any IP (code, design files, domain, hosting, analytics, repo) stay with them after final payment?YesWalk away. You'd be renting, not buying.
- Q5Did they trash your current site without looking at the codebase?YesWalk away. They're anchoring you toward a full rebuild.
- No red flags. Now weigh the green ones.
A real proposal requires understanding your business, your audience, your goals and the scope. An agency that drops a number after an intro call will either lose money and cut corners to compensate, or charge you for "scope changes" later.
The portfolio test is the most underused signal. Pull up three of their recent client sites on your phone right now. Then ask for names: many agencies have a senior consultant pitch and a junior developer execute. If the principal pitching you won't run your project, you're buying a different product than you think.
Be wary, too, of anyone who trashes your current site on day one. Sometimes a rebuild is right; often it isn't. The agencies worth hiring lead with honest assessment: "we'd need to audit before making that call."
The green flags worth weighting heavily
- A written process (discovery, design, build, QA, launch, post-launch) with deliverables and acceptance criteria for each phase.
- They've turned down work. The agencies that take every project fail most often.
- They explain why their tech stack fits your situation, not just that it's modern.
- A real care plan. Most site failures happen in months 2–12, not on launch day.
- Senior people in the room for discovery and design, and references you can actually call.
The 12 questions to ask in your first real meeting
- Can you walk me through your most recent client project, discovery to launch, with specific dates and people?
- Who specifically will be on my project, and what is each person's role?
- What does your typical timeline look like for a project this size, and what causes timelines to slip?
- How do you handle scope changes mid-project? Show me the change-order process.
- What's your payment milestone structure? What deliverables trigger each payment?
- What happens if I'm not happy with the first round of design?
- How do you handle SEO, structured data and performance during the build?
- What's your accessibility standard, and how do you verify it?
- What does post-launch support look like, for how long, and at what cost?
- Who owns the source code, design files, domain and hosting accounts at the end?
- What's your communication cadence and response-time commitment?
- Can I speak with two clients you've worked with in the last 12 months?
You don't need all twelve in one meeting, but get answers to all of them before you sign.
The contract clauses people forget
Most clients sign whatever standard agreement the agency hands over. Insist on these.
- IP transfer on final payment
- Written scope and acceptance criteria
- A change-order process with costs approved in writing
- 30–90 days of bug fixes included after launch
- The full source repo, not just a hosted site
- Hosting, domain and analytics in your name
- A termination clause that defines your exit
- Confidentiality and data-handling terms
IP transfer matters most: all source code, design files, content and credentials become yours when the final invoice is paid. The termination clause comes second. If the project goes sideways, define the exit (a refund, partial deliverables) before you sign, not after. Confidentiality terms matter most in regulated industries.
How to read between the lines of a proposal
A proposal tells you almost everything if you read it carefully. Here's what I look for.
| Feature | Good sign | Warning sign |
|---|---|---|
| Pricing | Line items | Lump sum you can't break down |
| Deliverables | Homepage design, 3 revisions | A stunning visual experience |
| Timeline | Tied to when you deliver content and approvals | 8 weeks, no dependencies |
| Assumptions | Listed clearly | Missing |
| Out of scope | Spelled out | Missing, so you'll pay for it later |
A lump sum can be fine for fixed-scope work, but you should always be able to ask what one piece costs on its own. And a line like "we assume you provide content by the 15th" protects both sides.
The cost of getting it wrong
Here's what clients I've onboarded after a failed first attempt typically went through: a slow, buggy site that didn't match the brand or didn't work, after 3–9 months of delay.
| Item | Range |
|---|---|
| Initial cheap quote | $4k–$12k |
| Remediation or restart | $15k–$40k |
| Total for the failed path | $19k–$52k |
| Done right the first time | $15k–$30k |
The lesson isn't "always go expensive." If one quote is dramatically lower than the others, understand specifically why before assuming it's a deal.
Freelancer, studio or agency?
- Right for small
- well-scoped projects
- Single point of failure
- No bench depth
- Senior execution
- Transparent process
- Manageable cost
- Capacity for large or multi-channel work
- You pay for process layers
I wrote a separate post on this for South Florida, but the short version applies anywhere: match the size of the team to the size and risk of the project.
In the end, picking an agency is a trust decision wrapped in a procurement process. The framework makes it a structured evaluation, but the question underneath stays the same: do I trust these people to do what they say, and to be honest with me when something goes wrong?
Want an honest read on whether we're the right fit? We'll tell you if we're not. Book a free 30-minute call, or see how we approach custom web development and what our pricing structure looks like.
Frequently asked questions
Should I work with a local agency or remote?
Local matters more for relationship-driven industries like law, real estate and healthcare, and for photography or on-site discovery. For pure execution remote works, though time zones matter at launch.
How do I verify an agency's portfolio is really theirs?
Check launch dates on the Wayback Machine, look up the team's individual portfolios, and call one of the listed clients.
How long should the evaluation process take?
Two to four weeks: an intro call, a discovery meeting, a proposal review and a negotiation pass. Being rushed to sign in 48 hours is a tell.
Can I get a refund if the project goes badly?
Only if the contract says so. Insist on milestone payments tied to deliverables, so a stalled project doesn't keep getting invoiced.
