How to Know Your Idea is Legit: Validation, Failure & Shipping What Sticks | StrictlyVC Athens 2026
TL;DR
Finny co-founder Victoria and Defraction CEO Johannes Galatanos share how founders can validate startup ideas by building self-trust through incremental risks, identifying urgent 'hair on fire' customer problems that drive immediate payment, and distinguishing genuine demand from polite encouragement.
🧠 Building Founder Confidence 2 insights
Accumulate small wins to flex risk-taking muscle
Victoria built confidence by progressively stepping out of her comfort zone, such as studying computer science at Stanford despite no prior programming background, proving she could achieve set goals before attempting to build a company from scratch.
Cross-industry exploration reduces fear of failure
Johannes developed risk tolerance by living in five countries and working across six different industries, making it intrinsic to start new ventures and pivot careers, including attending MIT mid-career to study quantum technologies.
🔥 Validating Ideas Through Customer Signals 3 insights
Identify 'hair on fire' problems with immediate willingness to pay
Victoria emphasized that the strongest validation is when customers urgently demand a solution, illustrated by Finny's first customer offering to pay double for early access to a product that didn't yet exist.
Apply the 'Mom Test' to avoid polite lies
Founders should ask customers about current pain levels and existing workarounds rather than hypothetical questions like 'would you use this,' as people naturally offer polite encouragement that masks true lack of interest.
Find customers with no alternative solutions
For deep tech, Johannes advised finding niche customers who genuinely have no other way to solve their problem and will tolerate early usability issues, validating the technology through letters of intent or initial contracts rather than opinions.
🚀 Product Evolution & Technical Relevance 3 insights
Maintain the problem while pivoting the solution
Victoria explained that while Finny's core problem—helping financial advisors grow—remained constant, the solution evolved from a static database to an automated outbound agent based on how customers actually used the product.
Bet on team velocity for fast-moving tech
To combat rapid AI obsolescence, Victoria advises selling customers on the founding team's commitment to staying at the forefront, as the product seen today may differ completely from the version delivered next month.
Target 100-1000x improvements in unsolved niches
Johannes noted that deep tech success requires finding applications where novel technology provides orders-of-magnitude improvements that conventional methods cannot achieve, such as quantum cameras processing 20,000 units per second versus one per hour.
⚠️ Learning from Failures & Funding 3 insights
Abandon ideas with lukewarm feedback
Victoria's initial idea for a QA testing agent failed because customers offered only vague interest, whereas Finny generated enthusiastic, immediate demand, revealing that polite responses are anti-signals for product-market fit.
Avoid regulatory-heavy initial markets
Johannes shared failures in microscopy due to lengthy regulatory timelines and semiconductor manufacturing due to technical mismatches, emphasizing the importance of finding the right beachhead application through rapid experimentation.
Prioritize private funding before government grants
Johannes recommended reversing the funding sequence, securing private capital first while treating government grants as a parallel stream, since bureaucratic delays and IP restrictions can stall early momentum.
Bottom Line
Stop asking customers if they like your idea and instead verify they have an urgent unsolved problem evidenced by hacked-together workflows and immediate willingness to pay.
More from TechCrunch
View all
Everyone from OpenAI to SpaceX is building their own chips | Equity Podcast
Major tech companies from OpenAI to SpaceX are developing custom AI chips to reduce costs and dependence on Nvidia, while the industry grapples with memory shortages and pivots toward 'neocloud' compute rental models amid skepticism about futuristic orbital data centers.
Is the US government's Anthropic ban accidentally helping the brand? | Equity Podcast
The US government banned Anthropic's Fable 5 and Mythos 5 AI models citing national security concerns, but the conflict may paradoxically boost the company's brand as both technologically superior and resistant to political pressure, while the UK announced sweeping social media restrictions for teenagers.
NEA’s Tiffany Luck says enterprises are still figuring out their AI ROI | Equity Podcast
NEA partner Tiffany Luck explains that while consumer AI approaches 'magic moments' with personal agents capable of managing mental load, enterprises are struggling to measure ROI amid runaway token spend, as the industry anticipates major IPOs from OpenAI and Anthropic that will test the sustainability of AI unit economics.
SpaceX, Anthropic, and OpenAI’s hot IPO summer | Equity Podcast
As SpaceX prepares for its historic IPO, the company has pivoted to leasing excess AI compute capacity to Google and Anthropic for nearly $1 billion monthly combined, while Waymo acquired Apple's abandoned Arizona proving ground for $220 million to expand autonomous vehicle testing infrastructure.