In established companies, the decision to build is usually made long before anyone has evidence that it should be. Roadmaps fill with client requests, executive conviction and competitive gap-fills, and the cost of a wrong bet hides below the waterline: years of maintenance, capacity pulled from the core business, and quiet write-offs. Drawing on his book Known Before Built, Andreas shows how leadership teams move from belief to evidence before committing serious money, using observed customer behavior, cheap assumption tests and pilots designed to end in a clear decision. He closes with the question every product leader hears, “Why can’t we just ship it?”, and how to answer it with a number.
Why the build cost is the cheapest part of a wrong decision, and the four cost layers (implementation, carrying, displacement, correction) that never appear in the business case.
A staged path from first signal to investment decision, with clear decision rights at each gate: continue, hold, pivot or kill.
How to answer “Why can’t we just ship it?” with a number instead of an opinion.
Audience
Executive teams, CEOs and COOs of industrial, consulting and professional-services firms, CPOs, executive roles in product, innovation and transformation, investment committees.
Lengths
Keynote
30 or 45 min
Executive session
90 min
Guest lecture
60–90 min
Abstract
Boards are funding AI on a simple promise: engineers ship faster, so the business grows faster. The first half may come true. The second does not follow. In a 2025 randomized trial by METR, experienced developers using AI tools took 19% longer on real tasks, yet believed they had been 20% faster. And where speed is real, it multiplies whatever enters the pipeline: a backlog of sales requests and stakeholder wishes becomes more features, shipped sooner, maintained forever. Andreas argues that the bottleneck was never engineering speed but deciding what is worth building, and shows what leaders must change so AI speed turns into outcomes.
Why output metrics (velocity, commits, features shipped) mislead boards about AI returns, and which outcome measures to track instead.
The electric-motor lesson from economist Paul David: gains come from redesigning the system around a tool, not from the tool.
Why AI removes engineering as product’s alibi, and what that means for accountability in the leadership team.
Audience
Leadership, executives, sponsors and budget holders for AI programs; stages for AI, digital transformation, innovation.
Lengths
Keynote
20, 30 or 45 min
Fireside
30–45 min
Panel
45–60 min
Abstract
Most customer research asks people what they think. Would they use it? would they pay? And treats the answers as evidence. They are not. People are unreliable reporters of their own behavior: when Porter and Nohria tracked 27 CEOs minute by minute, their time went somewhere very different from where they believed it went. This session teaches research built on past and observed behavior: interviews that surface real workflows, workarounds and spend; how to tell normalized pain from complaints nobody would pay to fix; and how to define, before the conversation, which behaviors would confirm or end a hypothesis. Participants leave with question patterns they can use the next morning. The session is the hands-on companion to Known before built.
The evidence hierarchy: reported behavior beats opinion, and observed behavior beats both.
Interview question patterns that pull out past behavior, workarounds and spend instead of predictions.
How to set confirming and disconfirming behaviors before the interview, so findings cannot be bent to fit the plan.
Audience
From COOs to Product managers, UX and market researchers, innovation and venture-building teams including corporate venture building, and the leaders who sign off on their research.
Lengths
Workshop
90 min or half day
Guest lecture
60–90 min
Abstract
In many companies, growth means hunting small incremental uplifts everywhere but the product itself. Onboarding, activation, adoption, retention, churn, expansion often have no clear owner once a product has reached a market. Andreas has spent more than fifteen years running experiments on both sides, from search, social and interface optimization to product experimentation led by behavioral goals. He argues that growth must work on the live product and feature development as much as on the marketing side. Growth runs its own experiments on what has shipped, alongside the product teams rather than downstream of them. A conversation for leaders who want experiments that move revenue, not only satisfaction and perceived usability upticks.
Why growth that stops at marketing leaves most revenue levers untouched: activation, pricing, adoption and retention live in the product.
How growth and product split the work after launch: product owns discovery and delivery, growth experiments on what is live, each with its own backlog.
How growth gets enough involvement in feature development to run product experiments without becoming a second product team.
Audience
CMOs, CROs, CPOs, VP/Director/Heads of growth, product and marketing; B2B, SaaS and growth conferences.
Lengths
Fireside
30–45 min
Panel
45–60 min
Formats
Keynote
20, 30 or 45 min
Fireside chat
30–45 min
Panel
45–60 min
As one of 3–5 panellists
Executive session
90 min
Leadership team, up to ~8
Workshop
90 min or half day
Up to ~30 participants
Guest lecture
60–90 min
MBA and executive education
All talks in English or German.
Biographies
Short
Andreas Mitschke has spent more than fifteen years building and leading at the intersection of product innovation and growth experimentation, a path that took him from UX design through code to growth in product and marketing. His work spans B2C and SaaS companies as well as B2B tech, industrial and deep tech organizations. He is the author of Known Before Built: How Established Enterprises Build Scalable Growth By Productizing Their Expertise.
Long
Andreas Mitschke has spent more than fifteen years building and leading at the intersection of product innovation and growth experimentation. His non-linear journey began at age 15. Before leaving his teens, he had run web design services, coded a Windows utility, and built a travel experience platform. From there, his path led from UX design to growth and product, across B2C and SaaS companies as well as B2B tech, industrial and deep tech organizations. He has advised startups and scale-ups through funds including 500 Global and Loyal VC, evaluated and brokered deal flow for venture funds across Europe, New York, and Silicon Valley, and helped build Germany's largest academic deep tech program. Closer to the ground, he coached founders at Mannheim's startup centers and lectured on entrepreneurship and product discovery at universities in Germany. He is the author of Known Before Built: How Established Enterprises Build Scalable Growth By Productizing Their Expertise.
Conversation angles
For podcasts, interviews and panels.
Why knowledge-rich companies struggle to turn their expertise into products
Why the consulting habit of saying yes to every client stalls productization
What “Why can’t we just ship it?” really costs
Why faster engineering with AI hasn’t produced better products
What to ask customers instead of their opinion
Quotable
From the book.
“Opinions describe what people believe they would do. Behavior shows what they did. Reported behavior beats opinion, and observed behavior beats both.”
Andreas Mitschke, Known Before Built
“Discovery is cheap. Correction is expensive.”
Andreas Mitschke, Known Before Built
“They are the rock stars. You are their fan.”
Andreas Mitschke, Known Before Built
“AI has made building cheaper. It has not made finding the right thing to build easier.”
Andreas Mitschke, Known Before Built
“The team keeps building. The product stops growing.”
Andreas Mitschke, Known Before Built
“The feature factory is running faster, but it remains a feature factory.”
Andreas Mitschke, Known Before Built
“We cannot ask our way to behavioral truth. We can only observe it.”
Andreas Mitschke, Known Before Built
“The status quo wins not because it is so good, but because switching is expensive in ways that do not appear on a pricing page.”
Andreas Mitschke, Known Before Built
“Workarounds don’t lie.”
Andreas Mitschke, Known Before Built
“An organization where killing a hypothesis carries career risk is an organization where bad ideas survive to production.”
Andreas Mitschke, Known Before Built
“When did your organization last kill a product on evidence and call it a success?”
Andreas Mitschke, Known Before Built
01 / 11
Press kit.
Portrait, square1920 × 1920 px, for programmes, posters and articlesPlaceholder
Portrait, round cut-outTransparent background, for speaker grids and event pagesPlaceholder