What is the single biggest red flag when hiring an MVP agency?

Hourly billing on a defined MVP scope. Real fixed-price agencies absorb scope risk; hourly agencies transfer it to you. An hourly engagement on a 2-week MVP typically becomes a 6-week engagement once the meter starts running. The fix is to require a fixed-price quote in writing before signing anything, with explicit inclusions, exclusions, and acceptance criteria.

How can I tell if an agency will hit me with change orders?

Three signs. First, the original scope document is vague or has no exclusions list. Second, the contract has an hourly add-on rate clause for 'additional work.' Third, the agency cannot show case studies where the original quote matched the final invoice. Any one of these means change-order inflation is the business model.

Should I worry if an agency does not have public case studies?

Yes, with nuance. Many agencies have NDAs that prevent public case studies, which is legitimate. But every legitimate agency can show you at least one case study under NDA in a conversation, with live URLs and architecture details. If they cannot even discuss past work in detail without giving away client names, they probably have not shipped real work.

What does code ownership mean in practice?

Code ownership means the codebase lives in your GitHub from day one, with you as the owner and the agency as collaborator. Design files (Figma, Sketch) in your account. Database in your cloud account. API keys provisioned by you. If any of these are held by the agency 'for convenience,' the agency has leverage to charge ongoing fees indefinitely. Walk if they will not commit to day-one ownership.

Are all hourly engagements bad?

No. Hourly is appropriate for genuine ongoing work where neither side can estimate scope upfront: research projects, exploration phases, ongoing maintenance. But an MVP is a defined scope. If a vendor insists on hourly for a defined MVP, the implicit reason is that hourly billing is more profitable for them, which means the project will take longer than necessary.

What if the agency seems great but does not include evaluation infrastructure?

Push back. For an AI MVP specifically, evaluation infrastructure (regression tests, output validation, cost monitoring) is not optional. Without it, the AI product will silently degrade when models update, fail unexpectedly on edge cases, and burn API budget without alerting. Either the agency adds it as included scope, or you walk. Building it later costs more than building it once at the start.

How do I walk away from an agency without burning bridges?

Be direct and specific. 'Based on our evaluation framework, your approach to [pricing model / code ownership / evaluation infrastructure] does not match our requirements. Thank you for the time.' Most agencies are professional enough to accept this. If they push back aggressively or try to renegotiate after you say no, the aggressive sales behavior is itself a red flag confirming the walk was correct.

MVP Agency Red Flags 2026: 12 Warning Signs Before You Sign

TL;DR

12 concrete red flags separate professional MVP development agencies from rate-card billing operations. The single biggest red flag is hourly billing dressed as fixed-price (vague scope + hourly add-on clauses = change-order inflation). Other deal-breakers: no published pricing, code held on agency infrastructure, no evaluation infrastructure for AI MVPs, scope documents without acceptance criteria, missing case studies, and 4+ week discovery phases for 2-week builds. Any single red flag from the list below is enough to walk. Founders who walk away cleanly save themselves the 3-6 months of pain that comes from signing with the wrong vendor.

Get a fixed-price scope from HouseofMVPs in 24 hours, all 12 red flags accounted for →

Why this matters

Most founders who get burned by an MVP agency do not get burned because the agency was incompetent. They get burned because the business model of the agency was misaligned with the founder's needs, and the misalignment was knowable before signing. The red flags below are signals you can spot in the first conversation, before any contract is signed, before any money changes hands.

The cost of missing these signals is not just the agency fee. It is months of lost time, broken or incomplete code, scope creep that triples the original quote, and the trust cost when investors or customers see a product that does not work as promised. The fix is to look for the warning signs upfront and walk if you see them, even if the agency seems otherwise impressive.

Red Flag #1: No published pricing on their website

How it manifests: The pricing page says "Contact us for a quote" or "Pricing varies by project." There is no specific dollar amount anywhere on the site.

Why it is a red flag: Real fixed-price agencies have published pricing because they can. Their delivery model is consistent enough to publish numbers. Agencies that hide pricing do so because the pricing changes based on what they think they can charge a specific prospect, which is rate-card billing behavior.

The walk-away script: "Your website does not publish pricing, which signals hourly billing in disguise. We need a fixed-price quote we can compare against other vendors. If you can send a fixed-price proposal in 48 hours, we will evaluate it. Otherwise we will move on."

Red Flag #2: Hourly billing for a defined MVP scope

How it manifests: The proposal quotes an hourly rate ($60 to $200 per hour) and an "estimated number of hours" for the project. Total cost is presented as "approximately $X, billed weekly."

Why it is a red flag: Hourly billing for a defined MVP transfers all scope risk to you. If the build takes longer than estimated (and it will), you pay more. The agency has zero incentive to be efficient. A 14-day MVP scope typically becomes a 4 to 8 week engagement on hourly billing, at 2 to 3 times the original quote.

The walk-away script: "We are evaluating MVP agencies on fixed-price terms because that is the model that fits a defined-scope project. If you can quote a fixed price for the same scope, we will consider it. Otherwise we will move on."

Red Flag #3: Scope document is vague or has no exclusions list

How it manifests: The scope document is a one-page bullet list of features without acceptance criteria. There is no exclusions section. When you ask "is X included," the answer is "we will figure that out as we go."

Why it is a red flag: Vague scope is how change orders happen. Every feature you assumed was included but is not in writing becomes a change order. Without an exclusions list, the agency has discretion to bill anything outside their narrow interpretation of the scope as extra.

The walk-away script: "We need a scope document with explicit inclusions, exclusions, and acceptance criteria per feature before we can sign. If you can produce that in 48 hours, we will continue the conversation. If not, we are not a fit."

Red Flag #4: No case studies they can discuss in detail

How it manifests: When asked for case studies, the agency points to a generic "Our work" page with screenshots and vague outcomes. They cannot or will not walk through architecture decisions, code, or specific challenges on any past project.

Why it is a red flag: Real production engineering produces specific stories. Real agencies have at least 3 to 5 case studies they can discuss in detail, including the technical challenges and how they were solved. Agencies that cannot discuss past work specifically either have not shipped much or have NDA constraints across the board (which is itself unusual).

The walk-away script: "Can you walk me through the architecture decisions on a project you shipped in the last 6 months? Specific tradeoffs you made, technical challenges, how you handled production issues." If they cannot, walk.

Red Flag #5: Code held on agency infrastructure until launch

How it manifests: When you ask about code ownership, the answer is "we transfer the code on launch day" or "we host it on our infrastructure and hand it over when the project is complete."

Why it is a red flag: Delayed code transfer is lock-in. If the project goes sideways mid-build, you have no leverage because you do not have the code. Even if the project completes successfully, you cannot inspect work-in-progress, audit the code, or migrate away if needed.

The walk-away script: "We require code in our GitHub from day one of the engagement. We will give you collaborator access. If this is not possible with your model, we are not a fit."

Red Flag #6: Discovery phase is longer than the build phase

How it manifests: The agency proposes 4 weeks of discovery + design before any code is written, for a 2 to 3 week build. The discovery deliverables are slide decks, PRDs, and design mockups.

Why it is a red flag: 4-week discovery for a 2-week build is rate-card billing. The agency charges $30,000 for discovery and then $30,000 for the build, when the customer actually needed a $7,499 MVP. Real agencies compress discovery to 1 to 3 days for an MVP and start building immediately.

The walk-away script: "Our timeline does not support a 4-week discovery phase. We need to ship the MVP in 4 weeks total. Can you compress discovery to under a week and start building in week 2? If not, we are not a fit."

Red Flag #7: Change-order inflation in past projects

How it manifests: When you ask for case studies where the original quote matched the final invoice, the agency hedges. "Most projects come in close to the original quote." "We always discuss change orders with the client first."

Why it is a red flag: Vague answers about quote-vs-invoice alignment mean change orders are the business model. Real fixed-price agencies stick to their quotes because the scope was defined tightly enough that change orders are rare exceptions.

The walk-away script: "Can you show me 3 case studies where the original quote was less than 10 percent off the final invoice?" If they cannot produce specific examples, walk.

Red Flag #8: No evaluation infrastructure for AI MVPs

How it manifests: For an AI MVP build, the proposal does not mention regression tests, output validation, cost monitoring, or evaluation harness. When you ask about quality monitoring after launch, the answer is generic ("we use standard monitoring") or non-existent.

Why it is a red flag: Production AI without evaluation infrastructure fails silently. When the underlying model gets updated (which happens every few months), quality regressions go undetected until users complain. Output validation failures crash the product. Cost monitoring gaps mean runaway loops can burn $5,000 in a week without alerting.

The walk-away script: "For an AI MVP, we require eval harness, output validation, and cost monitoring as part of the base scope. If those are extra costs or out of scope, we are not a fit."

Red Flag #9: One model for every use case

How it manifests: When you ask which model the agency will use, the answer is the same regardless of use case. "We always use GPT-5.5." Or "Claude is best so we default to it." No reasoning about cost profile, latency, quality requirements, or specific fit.

Why it is a red flag: Different AI MVPs need different models. A coding agent uses Claude Opus 4.7. A high-volume document processor uses Gemini 3.1 Pro. A general-purpose chat assistant might use either GPT-5.5 or Claude. An agency that defaults to one model regardless of use case has not done enough production AI to reason about the fit.

The walk-away script: "What is your reasoning for picking [their default model] for our use case specifically? What other models did you consider and why did you rule them out?" If they cannot answer concretely, walk.

Red Flag #10: Proprietary internal frameworks or tooling

How it manifests: The agency mentions their "internal framework" or "proprietary toolkit" that they use on every project. When you ask if you will own this framework, the answer is "no, but it makes us faster."

Why it is a red flag: Proprietary internal tooling is a moat against the customer. When you want to take development in-house or hire another agency, the existing code is tightly coupled to the previous agency's internal libraries. Switching costs go from a knowledge transfer to a full rewrite.

The walk-away script: "We require the code to be built on standard open-source frameworks that any other engineer or agency could continue. If your delivery depends on proprietary internal tooling we do not own, we are not a fit."

Red Flag #11: 100% upfront payment

How it manifests: The payment terms require 100% upfront, with the explanation that "this is industry standard" or "this is how we manage cash flow."

Why it is a red flag: 100% upfront eliminates your leverage. If the project goes sideways, you cannot withhold payment. The standard for fixed-price MVPs is 50% upfront and 50% on delivery, or milestone-based payments for larger projects. Anything else is a leverage move.

The walk-away script: "Our payment terms are 50% upfront and 50% on delivery, with delivery defined by the acceptance criteria in the scope document. If this is not workable, we are not a fit."

Red Flag #12: Aggressive sales pressure or limited-time discount tactics

How it manifests: "If you sign by Friday we can give you 20% off." Or "We only have one MVP slot left this quarter." Or "We are running a special promotion just for you."

Why it is a red flag: Real agencies do not need pressure tactics because their work sells itself. Pressure tactics are how vendors with weak books push prospects to sign before they finish evaluation. The pressure exists because the vendor knows if you take another week to compare options, you will pick someone else.

The walk-away script: "We are taking 2 weeks to evaluate vendors carefully. If your pricing or availability is contingent on signing this week, that is a signal we should walk." Then walk. The pressure will get more aggressive if you stay engaged.

How to walk away cleanly

Walking away from a vendor is not personal. It is a business decision based on evidence. The clean walk-away script for any of these red flags:

"Based on our vendor evaluation framework, your [specific issue] does not match our requirements. Thank you for the time you have invested in this conversation. We wish you the best."

Most agencies will accept this professionally. If a vendor pushes back aggressively, tries to renegotiate after you say no, or sends multiple follow-up emails, the aggressive behavior is itself confirmation that the walk was correct.

You do not owe an explanation beyond the one sentence above. You especially do not owe a discussion about why your requirements are what they are.

The HouseofMVPs answer to each red flag

For prospects evaluating us against this framework, here is how we score:

Published pricing: Three tiers visible on the pricing page: Validate $3,999, Launch $7,499, Scale from $14,999.
Fixed-price: Always. Hourly billing on a defined MVP is misaligned incentive.
Scope document: Written before signing, with inclusions, exclusions, acceptance criteria.
Case studies: Several public, several under NDA but discussable in detail on a call.
Code ownership: Customer GitHub from day one. Always.
Discovery: 1 to 3 days for an MVP. Compressed by pre-built foundations.
Quote alignment: We can show case studies where original quote = final invoice.
Evaluation infrastructure: Output validation, regression tests, cost monitoring included in Launch tier and above.
Model selection: Reasoned per use case. Claude Opus 4.7 for coding/agentic, GPT-5.5 for general, Gemini 3.1 Pro for large-context.
No proprietary frameworks: Standard stack (Next.js, Hono, PostgreSQL, Drizzle, Resend, Polar, Railway, Vercel). Any other engineer can continue the work.
Payment terms: 50% upfront, 50% on delivery.
No pressure tactics: Take your time. We are here when you are ready.

Related guides

How to Choose an AI MVP Development Agency: 25-Question Framework — full evaluation playbook
Top AI MVP Development Agencies in 2026 — ranked comparison
Best Fixed-Price MVP Development Agencies 2026 — agencies that pass the fixed-price test
What Is Fixed-Price Development? — pricing model explained

Ready to work with an agency that fails zero red flags? Get a fixed-price scope from HouseofMVPs in 24 hours →

MVP Agency Red Flags in 2026: 12 Warning Signs Before You Sign