What Founders Should Ask Before Hiring a Software Agency

Agency interviews often fixate on rates and timelines. Those matter, but they are downstream of better questions about ownership, risk, staffing, and decision-making. Founders who only negotiate price frequently buy expensive ambiguity that later appears as missed launches, brittle handoffs, and surprise invoices.

This article gives you a founder-ready question bank, explains why each question matters, and shows how to interpret answers in RFP responses, live interviews, and reference calls. Teams evaluating partners like SquartUp should leave with clarity on who does the work, how uncertainty is handled, and what done means in writing.

Use these questions as a living script. Send a subset in advance for written answers, then spend live time probing inconsistencies. The goal is not to embarrass a vendor; it is to simulate how you will solve hard problems together when production is unstable or scope shifts mid-sprint.

Tough questions attract serious partners. Soft questions attract soft commitments. If a firm becomes defensive when you ask about staffing substitution or credential ownership, believe that signal before the kickoff champagne.

Ownership questions that protect your future self

Ask who will staff your project day to day. Will you work with senior engineers, or is the pitch team different from the delivery team? Transparency here predicts quality. Request profiles for the core pod and ask what percentage of each person’s week is reserved for your account versus shared across clients.

Discuss intellectual property, repositories, and credentials early. You should own the code, domains, design files, analytics properties, and cloud accounts. Partners like SquartUp treat this as non-negotiable. If an agency hosts everything under their organization for convenience, negotiate an exit and transfer plan before kickoff — including who pays for the migration hours.

Clarify design system ownership and component libraries. Marketing sites and product interfaces both suffer when tokens and design files remain locked in a vendor seat you will lose at contract end. The same applies to documentation wikis and runbooks.

• Who holds production credentials on day one and on day ninety?
• Are repositories created under our GitHub or GitLab organization?
• How are design source files, tokens, and assets transferred?
• What happens to staging data, secrets, and monitoring dashboards at contract end?
• Who owns third-party accounts such as CDN, email delivery, and error tracking?

Risk, estimation, and change control

Request a sample technical approach for one risky feature such as payments, multi-tenant permissions, data migrations, offline mobile sync, or complex pricing. Teams that think in edge cases, security, and observability usually deliver more reliable products. Listen for how they name unknowns instead of hiding them behind false precision.

Ask how they estimate. Story points without historical velocity are theater. Better answers reference similar past work, spike budgets, and confidence ranges — for example, three to five weeks at roughly seventy percent confidence. Also ask how change requests are handled after kickoff: options with trade-offs, not surprise invoices that appear after a silent agreement in chat.

Probe quality bars explicitly. What automated tests are mandatory before merge? How is accessibility checked? What browser or device matrix is in scope? Saying we do QA is not an answer; tools, owners, and exit criteria are.

1. How do you run discovery and which artifacts do we keep afterward?
2. What does a typical sprint demo include and who must attend?
3. Which environments exist and who is allowed to deploy to each?
4. What is the definition of done for a user story on this team?
5. How do you handle severity-one incidents during hypercare?
6. How are analytics and success metrics validated before launch marketing begins?

Commercial and legal questions founders skip too often

Beyond day rates, clarify payment schedules, kill fees, warranty windows, and subcontracting. Some firms subcontract large portions of delivery without disclosure. Require written approval before any subcontracted work touches your product or data.

Ask about insurance and compliance posture if you handle personal data. Even if you are not yet enterprise-regulated, partners who understand least-privilege access, audit trails, and data retention will save you awkward customer security questionnaires later.

Discuss publicity rights separately from delivery. Case study usage, logos, and nondisclosure agreements should not be conflated with intellectual property ownership of the software itself. You can be generous with marketing permission while remaining strict about code ownership.

Clarify currency, taxes, travel, and tooling passthroughs. Small ambiguities here create disproportionate distrust when the first invoice arrives.

How to run the interview itself

Send questions in advance for written answers, then use the live call for probing. Split time deliberately: roughly twenty minutes on team and ownership, twenty on technical risk, fifteen on commercial terms, and fifteen on culture and collaboration norms. Record with permission so stakeholders who could not attend can review without a biased verbal summary.

Score independently before group discussion. Anchoring bias is real — the first charismatic presenter should not win by default. Require each interviewer to write a recommendation with risks before talking as a group.

Include at least one reference call focused on a project that struggled. Recovery stories reveal character better than highlight reels. Ask how the agency behaved during invoice disputes or scope tension, not only during launch celebrations.

Scenario walkthrough: marketplace MVP

You want a two-sided marketplace MVP in four months. Strong agency answers will challenge scope: which side is constrained first, what trust and safety rules exist, how payments and disputes work, and which analytics prove product-market fit. Weak answers jump to screens and technology stacks without questioning the business model.

Give both shortlisted agencies the same brief and compare how they structure discovery spikes, risk registers, and phased releases. The better partner often proposes a narrower first release that learns faster — and can explain what learning would change phase two without rewriting everything.

Scenario walkthrough: internal ERP module

Your operations team wants inventory and purchasing in one system. Ask agencies how they map current workflows before writing code, how they handle dirty historical data, and how they train non-technical users. Adoption failure is more common than technical failure in internal tools.

Listen for talk of parallel runs, permission models, and reconciliation reports. If the conversation stays only in interface mockups, keep probing until operational reality appears.

Scenario walkthrough: replatform under existing traffic

You must migrate a content-heavy site that already ranks and converts. Ask about redirect strategy, content freeze windows, performance budgets, crawl validation, and rollback. Partners who treat search engine optimization as marketing’s problem after launch will cost you organic revenue and executive patience.

How answer quality differs by partner type

• Product-minded studios ask about users and metrics before quoting screens
• Staffing vendors ask about seat counts and duration before problem framing
• Sales-led agencies over-index on vision decks and under-specify staffing
• Engineering-led shops may skip change-management talk — probe adoption explicitly
• Offshore-only pitches that cannot describe overlap hours usually underestimate coordination cost

Interview pitfalls to avoid

• Letting price be the first filter before fit and staffing clarity
• Allowing only marketing stakeholders to attend technical interviews
• Accepting we will figure it out in sprint one for known high risks
• Skipping security and credential questions until late legal review
• Failing to document scores with evidence links for later leaders

Frequently asked questions

What if an agency refuses to share team names before contract?

Treat that as a serious risk signal. At minimum, require named roles and seniority bands with a substitution clause tied to your approval rights.

Should we ask for unpaid design samples?

Prefer paid, scoped exercises. Unpaid speculative work often selects for desperation or recycled templates, not thoughtful discovery.

How do we compare retainers versus fixed bids?

Match the model to uncertainty. High uncertainty favors retainers or time-and-materials with a capped first phase. Clear brochure sites can be fixed-bid with explicit assumptions.

What references should we call?

Ask for a client whose project struggled and recovered, not only a highlight-reel success. Also ask how the agency behaved during invoice disputes or scope tension.

How detailed should acceptance criteria be?

Detailed enough that a third engineer could verify them. Include performance, accessibility, analytics, and content entry checks — not only visual resemblance to mockups.

A note on internal readiness

Your internal readiness is part of the evaluation whether you admit it or not. If you cannot name a product owner with authority to decide, even an excellent agency will thrash. Nominate that owner before final vendor selection and give them time to participate in interviews.

Likewise, prepare access to analytics, brand guidelines, existing repositories, and known integration constraints. Agencies that ask for these early are not being difficult; they are trying to estimate honestly.

When you want a partner that answers ownership and risk questions without flinching, talk with SquartUp. The best interviews feel like the beginning of delivery collaboration, not a theater performance.

Finally, keep a parking lot of unanswered questions after each call. The firm that follows up with written clarifications usually behaves the same way during production incidents — and that habit is worth more than a slightly lower hourly rate.

Building a decision memo your board will respect

Translate interview scores into a short decision memo: problem framing, options considered, recommendation, residual risks, and the first ninety days of working norms. Boards and executive sponsors do not need every technical detail; they need confidence that you compared like with like and protected ownership.

Attach the common brief, the scorecard, and redacted proposal excerpts. When finance asks why you did not pick the cheapest bidder, the memo should answer before the meeting starts. That preparation also reduces second-guessing during early delivery friction.

Include a kill criteria section: what evidence would cause you to pause or retender. Healthy partnerships can survive hard gates. Vague optimism cannot. Partners such as SquartUp would rather lose a poorly fit deal than inherit an unclear mandate.

Revisit the memo after discovery. If reality diverges from sales promises, you have a documented baseline for course correction rather than a fog of memories.

Questions for reference calls that reveal truth

Ask references what they would hire the agency for again and what they would not. Ask how estimates compared to actuals on the riskiest feature. Ask who actually answered the phone during a production issue and how long remediation took.

Ask whether documentation and environments were usable by a new hire after the agency reduced involvement. The answer separates true handoff from ceremony. Also ask how change requests were negotiated when marketing suddenly needed a new template before a campaign.

If every reference sounds identically glowing, request one more from a longer-ago project. Time reveals maintainability. Yesterday’s launch party does not.

Negotiating without poisoning the relationship

Negotiation should clarify risk allocation, not squeeze every hour of goodwill out of the team you need motivated next month. Trade scope and sequencing before you trade quality. Cutting testing, documentation, or staging to hit a number is how expensive failures are born.

Be explicit about response times, holiday coverage, and who pays for third-party tools. Put substitution rules in writing: if a named lead leaves, you approve the replacement. These clauses feel stiff in sales and feel essential in week twelve.

Share your internal constraints honestly — budget ceilings, immovable launch dates, compliance reviews. Surprises late in procurement create adversarial energy that leaks into delivery.

Additional implementation notes

When you operationalize the ideas in this guide, assign a single accountable owner, set a ninety-day review date, and write down what evidence would count as success. Teams that skip ownership turn good advice into unread bookmarks. SquartUp can help facilitate that operating cadence when you want an external partner in the loop.

Revisit assumptions after the first release. Markets change, analytics reveal surprises, and constraints shift. Treat the article as a playbook to adapt, not a script to recite.

Practical next step 1: document the current baseline metrics related to this topic, pick one high-leverage improvement for the next two weeks, and schedule a demo where stakeholders see the before-and-after. Small visible wins fund larger roadmap investments and keep momentum honest. SquartUp encourages this incremental approach on client engagements because it reduces big-bang risk.

Practical next step 2: document the current baseline metrics related to this topic, pick one high-leverage improvement for the next two weeks, and schedule a demo where stakeholders see the before-and-after. Small visible wins fund larger roadmap investments and keep momentum honest. SquartUp encourages this incremental approach on client engagements because it reduces big-bang risk.

Practical next step 3: document the current baseline metrics related to this topic, pick one high-leverage improvement for the next two weeks, and schedule a demo where stakeholders see the before-and-after. Small visible wins fund larger roadmap investments and keep momentum honest. SquartUp encourages this incremental approach on client engagements because it reduces big-bang risk.

Key takeaways

• Who is on the delivery team and what is their seniority mix?
• How do you estimate and communicate risk with ranges?
• What does ownership of code and infrastructure look like in writing?
• How are change requests handled after kickoff?
• Can you walk through a risky feature approach including edge cases and rollback?

Bring this question bank to your next agency interview. Ready to put these ideas into practice? Visit squartup.com or contact the SquartUp team to scope your next launch.