I have sat through a lot of pitches. On both sides of the table. As a strategist presenting to clients, and as a client evaluating agencies. After 15+ years, I can tell you with absolute certainty that the pitch process is one of the most wasteful rituals in the advertising industry.
Not because the people are bad. Because the format is broken.
The numbers are damning
US agencies spend $12.5 billion a year on pitching. That is 17% of total industry revenue, according to Forrester. The average win rate across 850 agencies? 19%, per the Bedford Group's 2026 benchmarking study. Each failed pitch costs an agency roughly $28,400 in unbillable hours.
Clients consistently underestimate pitch costs by five to six times. And only 5.5% of brands contribute anything to pitch costs. The rest treat it as free labour.
These are not edge cases. This is the system working as designed. A system designed for a world where building things was expensive and the only way to demonstrate capability was to perform it.
The pitch tests performance. The client needs production.
Here is the fundamental problem. A pitch is a performance optimised for a room. Strategy theatre. Mood boards. Tissue sessions. Credential reels. The A-team shows up for 90 minutes, then the B-team shows up on day one.
The EACA found that six out of ten winning pitch ideas are never implemented. The pitch selects for ideas that win rooms, not ideas that survive contact with reality. That statistic alone should make every procurement team rethink how they evaluate agencies.
The cost of building just collapsed
Two years ago, building an MVP took 8 to 12 weeks and a team of three. Today, a senior practitioner with AI tools ships a working product in days, not weeks. The cost dropped from tens of thousands to hundreds.
This changes the calculus entirely. The reason pitches existed as theatre was because building the actual thing was too expensive to do speculatively. That excuse is gone. When you can build a working prototype for less than the catering budget of a pitch, the only reason to pitch with a deck is because you cannot build.
And that tells the client everything they need to know.
A prototype is harder to fake than a deck
Decks let you hide behind polish. A working prototype reveals what you actually understand about the problem, the user, and the constraints. It is evidence, not promise.
The micro-decisions that reveal whether someone truly understands a brand, an audience, a problem space – those only emerge when you are building. The choice of typeface. The error message copy. The loading state. These are invisible in a pitch deck. They are unavoidable in a prototype.
Chemistry happens under real conditions too. The whole point of a pitch is to test whether you can work together. But pitch conditions are artificial. Building something together, even something small, reveals how a team thinks, communicates, and solves problems. That is the real chemistry test.
Who is already doing this
This is not theoretical. IDEO won Apple's first mouse contract by building a working prototype, not presenting a deck. R/GA uses high-fidelity prototypes in pitches for Nike, Google, and Samsung, and calls it a competitive advantage in generating new opportunities. ustwo wins clients like Toyota through build sprints, not credentials reels.
Y Combinator's Demo Day requires working products, not decks. Droga5 moved to paid pitches only and their win rate went from roughly 20% to 60-70%. When both sides have skin in the game, the output is closer to reality.
Forrester's Jay Pattisall has been advocating for paid proof-of-concept collaborations since 2023. The Pitch Positive Pledge, launched by the IPA and ISBA in 2022, has 190+ signatories. The industry knows the current model is broken. It just has not had a viable alternative.
AI changes that.
What this looks like in practice
The new model is not complicated. Brief plus chemistry session: one day, meet the team, align on the problem, no credentials reel. Then a paid prototype sprint: one to two weeks, build a working version of the idea, client involved throughout. Present with a URL, not a PDF. Let stakeholders use it. Let it speak for itself. Then iterate together based on real feedback, not room dynamics.
Six weeks, not six months. Both sides paid. The output is a product, not a promise.
The pushback
I hear two objections. The first: pitches test strategic thinking. They do. But strategic thinking also shows in what you choose to build, what you leave out, and how you frame the problem the prototype solves. Strategy is not a separate deliverable. It is embedded in every decision the prototype makes visible.
The second: not every brief suits a prototype. True. Brand positioning, comms strategy, media planning – these need different formats. But for anything with a digital, product, or experience component, the prototype is closer to the work than the deck ever was.
From my own work
I stopped pitching with decks. I started pitching with products. Every project in my portfolio started as a prototype built to solve a real problem. 20+ live products. Not one of them began with a tissue session.
The strategy is in the build. The taste is in the details. The proof is in the URL.
The Pitch Positive Pledge was a start. It was not enough. Pledges change conversation. They do not change behaviour. What changes behaviour is making the alternative cheaper, faster, and better for everyone involved.
AI just made that possible. Stop performing. Start building.
The full 24-slide deck is at mikelitman.me/prototype.
