System Alphas

Build less, learn faster: what the research says about launching new products.

Most new products fail from building the wrong thing, not from building it badly, and AI is changing how fast they get built. We gathered findings from CB Insights, Pendo, McKinsey, GitHub and Stack Overflow, and added what each one means for your MVP.

5 sources2018–2026 research7 min readBy System Alphas
1. Fit beats features43% of failed startups cite poor product-market fit; most also ran out of cash first.
2. Most features go unused80% of features in the average product are rarely or never used.
3. AI speeds the build, with review84% of developers use AI tools and tasks finish up to 55% faster, but 46% distrust AI output.
CB Insights · 2026Pendo · 2019McKinsey · 2018GitHub · 2022Stack Overflow · 2025
Finding 01CB Insights · March 202643%of failed startups cite poor product-market fitWhy Startups Fail: Top 9 Reasons

What the research found

  • Analysis of 431 VC-backed companies that shut down since 2023; reasons identified for 385 of them.
  • 70% ran out of capital, which CB Insights calls the final cause of death, not the root problem.
  • 43% cite poor product-market fit, 29% bad timing and 19% unsustainable unit economics.

What it means for you

Most products die from building the wrong thing, then running out of money while finding out. Validate demand before you spend the budget.

How we apply it

Our strategy sprint tests the problem and the core feature set before any build.

See Product Strategy Read the source
CB Insights research report, March 5, 2026.
Finding 02Pendo · 201980%of features in the average software product are rarely or never usedThe 2019 Feature Adoption Report

What the research found

  • Based on product usage data across Pendo customers.
  • Publicly traded cloud software companies invested up to $29.5 billion developing those rarely used features.

What it means for you

Every feature you do not need costs money to build and maintain. A focused MVP ships the few features users actually use.

How we apply it

We scope the smallest feature set worth building, then add only what data supports.

See our process Read the source
Pendo Feature Adoption Report, 2019.
Finding 03McKinsey & Company · October 201832 ptsfaster revenue growth for top-quartile design performersThe business value of design

What the research found

  • Five-year study of 300 publicly listed companies across countries and industries.
  • Top-quartile McKinsey Design Index scorers grew revenue 32 percentage points faster than industry peers.
  • Their total returns to shareholders grew 56 percentage points faster.

What it means for you

Design is not decoration. Products that are easy to understand win more customers, even at MVP stage.

How we apply it

UX research and design are part of every MVP we build.

See UI/UX Design Read the source
McKinsey Quarterly, October 25, 2018.
Finding 04 AI effectGitHub · September 202255%faster task completion for developers using an AI pair programmerResearch: quantifying GitHub Copilot’s impact on developer productivity and happiness

What the research found

  • Controlled experiment with 95 professional developers writing an HTTP server in JavaScript.
  • The Copilot group finished in 1 hour 11 minutes on average, versus 2 hours 41 minutes without it.
  • 73% said it helped them stay in flow; 87% said it preserved mental effort on repetitive tasks.

What it means for you

AI-assisted engineering shortens the build without cutting corners, so you reach real users sooner.

How we apply it

Our engineers build with AI pair-coding, then review and test every line.

See AI-Built MVPs Read the source
GitHub Blog, September 7, 2022 (updated May 2024).
Finding 05 AI effectStack Overflow · July 202584%of developers use or plan to use AI tools, yet 46% distrust their accuracy2025 Developer Survey (reported by InfoWorld)

What the research found

  • Survey of more than 49,000 developers in 177 countries.
  • AI tool use rose from 76% in 2024 to 84% in 2025.
  • Distrust in the accuracy of AI output rose from 31% to 46%.

What it means for you

AI now builds a lot of first-version code, but unchecked AI output is a real risk. Speed only pays off with experienced review.

How we apply it

Every AI-generated line in your MVP is reviewed and tested by a senior engineer.

See AI-Built MVPs Read the source
InfoWorld, July 30, 2025, reporting the Stack Overflow 2025 Developer Survey.

Sources

  1. CB Insights, Why Startups Fail: Top 9 Reasons, March 2026. https://www.cbinsights.com/research/report/startup-failure-reasons-top/
  2. Pendo, The 2019 Feature Adoption Report, 2019. https://www.pendo.io/resources/the-2019-feature-adoption-report/
  3. McKinsey & Company, The business value of design, October 2018. https://www.mckinsey.com/capabilities/growth-marketing-and-sales/our-insights/the-business-value-of-design
  4. GitHub, Research: quantifying GitHub Copilot’s impact on developer productivity and happiness, September 2022. https://github.blog/news-insights/research/research-quantifying-github-copilots-impact-on-developer-productivity-and-happiness/
  5. Stack Overflow, 2025 Developer Survey (reported by InfoWorld), July 2025. https://www.infoworld.com/article/4031673/ai-use-among-software-developers-grows-but-trust-remains-an-issue-stack-overflow-survey.html

Figures are quoted from the publishers listed above; follow the links for full methodology. System Alphas is not affiliated with these organizations. Sections titled “What it means for you” and “How we apply it” are our own analysis.

Launch smarter

Test your idea before you bet the budget.

Our strategy sprint defines the smallest product worth building and a fixed launch date.