Overview of Practical Founders Podcast — Episode #184: Fixing The Software Development Mess For Non-Technical Founders (guest: Keith Shields)
This episode features Keith Shields, co-founder and CEO of Designly, in conversation with Greg Head. They discuss the persistent product-development problems that non-technical founders face, how Designly evolved from those early failures into a disciplined outsourced engineering partner, and practical ways founders can get unstuck — from audits and prototype validation to team composition, process rituals, and judicious use of AI.
Key takeaways
- Non-technical founders repeatedly fall victim to "black box" engineering: poor communication, unclear scope, missed expectations, and buggy deliveries.
- Designly’s remedy: multidisciplinary, full-time-assigned teams + strong communication rituals and an early two-week Solution Lab to align vision and scope.
- Start fast and validate: get to your first paying customer quickly (velocity to first dollar) — iterate from there.
- If you already have a messy product, get an outside audit (an “impact week”) — it’s often worth paying for objective feedback before making a risky change.
- Use hypothesis-driven development: define business metrics for each feature, instrument them, build the smallest implementation that tests the hypothesis, and measure.
Guest & company snapshot
- Guest: Keith Shields — co-founder & CEO, Designly.
- Company: Designly (designli.co) — a nearshore custom software development service focused on non-technical founders and vertical SaaS.
- Track record: helped ~200–250+ products over 12 years; ~94 employees (nearshore in Latin America — El Salvador, Colombia, Argentina); time-zone aligned to U.S. Central/Eastern.
- Core audience: bootstrapped/practical founders, friends & family-funded startups, vertical SaaS founders.
Problems discussed (common founder pain points)
- Agencies or freelancers operate as a black box — little to no proactive communication.
- Fractional teams across many clients deliver low velocity and low ownership.
- Delivered features are late, buggy, or different from expectations.
- Founders don’t know what to prioritize; they build features users don’t pay for.
- Bad code foundations (e.g., poor structure, exposed secrets) from quick validations can become expensive technical debt.
- Difficulty finding engineers who are both technically competent and customer-focused.
Designly’s approach / solution
- Founding ethos: remove the “pain” and give founders peace of mind through transparent communication and treating clients as partners.
- Core principles:
- Full-time developer assignments (no rotating, fractional devs).
- Multidisciplinary teams (design, engineering, product owner, QA).
- Ritualized communication (sprint planning, demos, monthly strategy).
- Hypothesis-driven product decisions and measurement-first mindset.
- Offerings:
- Solution Lab: 2-week design sprint (about 160 hours) that produces a clickable Figma prototype + development plan, estimates, and roadmap. Money-back guarantee if not satisfied.
- Impact Week: an audit (code, design system, project management) to diagnose problems — used for recovery from bad projects.
- Code takeovers & dedicated development teams for ongoing work and scaling.
Services, typical team structures & pricing
- Typical Designly team (example):
- Full-time front-end engineer
- Full-time back-end engineer
- Half-time tech lead (also contributes)
- Half-time product owner (20 hours/wk; PO typically handles 2 clients)
- Designer (≈10 hrs/wk)
- QA (half-time)
- Typical monthly cost: common teams often fall in the $30k–$33k/month range.
- Lean/early team alternative: single full-stack + tech lead + product owner — roughly $12k–$15k/month.
- Team scale: teams usually start 5–6 people and can expand to ~14–15 for larger initiatives; Designly can scale “accordion-style” for bursts of work.
Process & rituals (how they actually work together)
- Solution Lab first (when starting from scratch) to align vision, create prototypes, and size features into a roadmap.
- Impact Week audit for messy/existing builds (no-cost diagnostic).
- Post-audit takeover process: ~3 weeks to set up local dev environments, fix low-hanging issues, and get the team productive; founder should expect ~1 month of ramp.
- Regular meetings and artifacts:
- Sprint planning (Monday) — commit to tasks for the two-week increment.
- Sprint demo & QA review (Friday) — show completed work and test evidence.
- Monthly strategy review — prioritize hypotheses and roadmap 1–3 months out.
- Direct Slack/JIRA or Basecamp communication; the whole pod is accessible, not hidden behind an account manager.
- Emphasis on acceptance criteria, test-case documentation, and transparency about what was tested and shipped.
Validation, metrics & hypothesis-driven development
- Every new feature should map to a measurable business metric (e.g., conversion, retention, ARPU).
- Instrumentation and tracking (event logging/analytics) must be built alongside the feature to know whether it worked.
- Start with the smallest implementation that tests the hypothesis (prototype → minimal dev → measure → iterate).
AI: how Designly uses it and recommendations
- Practical adoption:
- Developers use GitHub Copilot and other tools to write code faster.
- AI speeds coding and testing but can also produce bad code faster; systems thinking and review remain critical.
- For AI-first features:
- Don’t “sprinkle” AI just to be trendy — tie it to a business outcome.
- Typical AI integration patterns: chat interfaces that sit on top of deterministic systems; orchestration & API design are key.
- For AI-heavy projects, Solution Lab work may focus on system integration diagrams (data flows, orchestration) rather than UI prototypes.
- Advice: start small, validate, instrument, and don't replace crucial deterministic workflows with probabilistic agents prematurely.
Notable quotes
- "Velocity to your first dollar — how quick can you get your first dollar? That's the mission and nothing else matters until you accomplish that mission."
- "Trust your gut and get outside perspective." (On hiring audits/impact weeks)
Concrete advice & action items for founders (checklist)
- If you’re pre-launch:
- Run a lean validation (prototype/“lovable” front-end) and aim for your first paying customer quickly.
- Use a Solution Lab or equivalent to convert ideas into a testable prototype and a scoped development plan.
- If you already have a messy product:
- Trust your instincts. Commission an outside audit (code + design + PM process). Pay for it if needed — it’s often worth it.
- Consider an Impact Week or similar diagnostic before making sweeping changes.
- Teaming and process:
- Prefer dedicated, full-time-assigned developers to fractional-hour models.
- Use multidisciplinary pods (design + PO + dev + QA).
- Put product rituals on the calendar (planning, demos, monthly strategy).
- Implement hypothesis-driven development and instrument every feature with measurable metrics.
- With AI:
- Use AI tools to accelerate work (Copilot, automated testing) but maintain strong code review and architectural discipline.
- Add AI features only when tied to a clear business outcome; validate incrementally.
Useful numbers (from the episode)
- Designly Solution Lab: 2 weeks / ~160 hours (money-back guarantee).
- Solution Lab cadence: about 1 per week; ~50 Solution Labs/year.
- Conversion: ~36% of Solution Lab customers convert to a development engagement.
- Typical Development Team cost: ~$30k–$33k/month; lean teams ~$12k–$15k/month.
- Team size at Designly: ~94 employees; nearshore across Latin America.
Final summary
This episode is a practical field guide for non-technical founders who need a reliable path from idea or broken MVP to a maintainable, measurable product. The core prescription: validate fast, get an aligned multidisciplinary team dedicated full-time, create transparent rituals and documentation, measure hypotheses, and only then scale — using AI as an accelerator, not a substitute for systems thinking. If you’re stuck, start with an audit (impact week) and a two-week alignment sprint (Solution Lab) before you commit more runway.