Why the UK has spent thirty years running the one mechanism Colin Mason's evidence says does not work. And what a programme designed for how angels actually decide looks like instead.
There is a particular kind of British invention – the pitch competition – that has achieved, over thirty years, an almost perfect record of doing the opposite of what it was designed to do.
Pitch competitions were invented to reduce the information friction between founders and investors. They promised to compress the discovery problem into a single standardised format, bring multiple investors into the room at once, and give promising companies a route to capital that did not require a decade of City networking. The theory was elegant. The execution is a disaster.
A generation later, the United Kingdom runs pitch competitions, demo days, and accelerator showcases at industrial scale, and the measurable result is that roughly 90% of early-stage funding proposals are rejected in under twenty minutes. The average founder attending the average pitch event is rejected by the average angel before either of them has had time to finish a coffee. The pitch competition is the most efficient rejection machine ever built.
The puzzling thing (and this is the point of this essay) is that we keep building more of them.
The Mason Problem
Colin Mason is a professor at the University of Glasgow who has spent thirty years studying how UK angel investors actually make decisions. If you want to understand why the pitch competition fails at its stated goal, his research is the place to start. A handful of findings from his work, replicated across multiple studies and three decades, have become so well-established that they ought to be design premises for any serious piece of investment infrastructure.
The first is that angels decide in under twenty minutes, then rationalise on substance. Colin Clark, working in the Mason tradition, observed twenty-four angels rating live pitches on thirty-two factors and found that presentational factors (clarity, structure, confidence, handling of questions) correlated more strongly with investor intention than market, product, or financial factors. The angels, asked to explain their decisions, gave reasons drawn almost entirely from substance. They decided on style; they reported deciding on merit. Neither they, nor the founders they rejected, had any reliable way of telling the difference.
The second is that the decision is dominated by the founder, not the product. In a 2017 survey of 238 UK angels, Mason and his colleagues found 59% citing “people” as their top rejection category, ahead of product, market, financials, and business plan. In thirty accompanying interviews, twenty-seven of thirty angels (90%) said the same. Of the specific sub-factors within the people category, the single largest was not competence, not track record, not domain expertise. It was honesty. Founders hide their weaknesses instinctively, imagining that concealment protects them. Mason’s evidence says the opposite. The founder who volunteers a risk first is not revealing a flaw. They are proving they are the kind of person who can be invested in for seven years without surprises.
The third finding, and in some ways the most consequential, is that the decision is non-compensatory. Angels do not add up your strengths, subtract your weaknesses, and produce a weighted score. They eliminate. A single fatal flaw in any dimension causes rejection regardless of how strong everything else is. Maxwell, Jeffrey and Lévesque, analysing 150 episodes of Dragons’ Den, found that 89% of rejections cited one specific deal-killer. A brilliant founder with a missing IP assignment is rejected. A merely adequate founder with no missing fatal flaws survives. Revenue does not compensate for a messy cap table. Traction does not compensate for co-founder equity confusion. The border guard, as I have argued elsewhere, is not interested in your personality.
Read these three findings together and you arrive at an uncomfortable conclusion. The pitch competition (the ten-to-twenty minute presentation in front of an unfamiliar investor) is the single worst conceivable mechanism for making the decision angels are actually trying to make. It puts a twelve-month judgement into a twelve-minute window. It forces the founder to perform rather than behave. It rewards presentation polish, which angels cannot consciously detect, and punishes honest disclosure, which they consciously respect. It produces a 90% false-negative rate.
And yet we keep building more of them.
The Sutherland Observation
Rory Sutherland’s general thesis (the one he has been patiently restating in various forms for most of his career) is that a surprising number of commercial problems are design problems masquerading as other kinds of problems. People do not need better incentives, they need a different menu. Trains are not slow, they are boring. Capital is not scarce, it is illegible. The solution to most market failures is not more of the existing mechanism. It is a different mechanism.
The UK equity gap is a case in point. It is standardly described as a capital problem – there isn’t enough money for early-stage founders. Mason’s evidence, read in full, says this is not quite right. There is plenty of money. There are approximately twenty thousand active angels in the UK, writing roughly one-point-five billion pounds a year. What there isn’t is a mechanism for matching that money to founders in a way that allows the actual decision to be made well. The ten-minute pitch is the bottleneck. Every additional pitch event, every better-designed demo day, every more-polished accelerator showcase, is an investment in making the bottleneck more efficient at doing the thing the evidence says should not be done.
A design-led response would not try to build a better pitch competition. It would ask a different question: if the decision cannot be made well in twenty minutes, what does a mechanism that gives the angel long enough to see the founder clearly look like?
The £100,000 Startup Race
This is the question the £100,000 Startup Race is designed around.
The Race is a twelve-month competition in which fifty founders, each already trading at £5,000 in annual recurring revenue, publicly race to ~£100,000. Bank-verified revenue is the primary leaderboard metric. The winner receives a £100,000 equity investment prize. So far, so familiar. Every good UK accelerator does something in this general shape.
What makes the Startup Race different is what sits around the revenue race.
First, there is a set of milestones (we call it the Investment Readiness Constellation) designed around Mason’s non-compensatory finding. Each milestone eliminates one of the specific, evidenced deal-killers that angels reject on. A clean cap table. IP formally assigned to the company. A founders’ agreement with vesting. SEIS Advance Assurance. Documented customer evidence. A valuation narrative the founder can defend. A weakness-disclosure document that pre-empts the honesty test. A recorded Q&A stress test with real angels. Each milestone exists because Mason’s evidence, or the closely adjacent literature, says angels kill deals on its absence.
The Constellation is not a set of nice-to-haves. It is a programme of fatal-flaw inoculation. The racer who completes it has removed themselves from the elimination-by-aspects screen on which 90% of proposals fail. That is a specific, defensible claim.
Second, there is an observing network of angel investors who watch the twelve months of the Race in real time. This is the structural innovation that turns the mechanism around. The standard pitch competition asks an angel to make a twelve-month judgement in twelve minutes.
The Startup Race does the reverse: it gives the angel twelve months of observable founder behaviour before any decision need be made. How does the founder respond to a bad week? How do they handle their first unhappy customer? When they say they will ship something, do they ship it? When they disclose a weakness, is it followed by action? These are the questions angels care about (Mason’s research says they care about them more than any pitch content) and they are questions a twelve-minute format cannot answer but a twelve-month one can.
Third, there is a mechanism we call the Golden Buzzer. At any point during the Race, an observing angel in the network can make an investment offer to a racer on their own conviction. The offer must economically dominate the £100,000 prize – either a super angel committing at least that amount on at least as good terms, or a syndicate round of £150,000 to £250,000 led by a champion angel.
A Startup Racer who accepts exits the revenue race and becomes a Class of 2026 Alumnus with ongoing visibility to the network. Between two and four buzzes happen per cohort. Timing is unpredictable. Buzzes are public. The criteria are Constellation completion and growth trajectory, not absolute revenue rank.
The buzzer exists because Mason’s group-dynamics research identifies the trusted-intermediary referral as the single most powerful shortcut in the angel decision process, and because probability-weighting research (Tversky and Kahneman 1992) shows a small, visible, time-unpredictable chance of a large positive outcome keeps the mechanism psychologically live across the full twelve months for every racer, not just the leaderboard top ten.
Fourth, there is a graduation architecture for the forty-nine founders who will not win the £100,000. Every completing racer receives a Week 51 investor memo, durable alumnus status, inclusion in a six-month follow-up programme designed to capture the stories of post-Race fundraising wins, and twelve months of warm-introduction rights into the observing angel network after the Race ends.
This matters because the zero-sum nature of the revenue prize would otherwise produce a motivational cliff at month four, when the non-top-ten racers realise they cannot win and start to disengage. The graduation architecture converts the Race from a lottery into a programme that materially improves the fundraising position of every founder who finishes it; regardless of where they finish.
None of these elements is, individually, novel. Accelerators have milestones. Angel networks have intros. Some programmes have alumni.
What makes the Startup Race structurally different is the integration: twelve months of longitudinal observation, a milestone structure tied explicitly to Mason-evidenced deal-killers, a trusted-intermediary investment mechanism, and a graduation architecture that makes the year valuable to all fifty participants. Every element responds to a specific finding in Mason’s thirty-year research programme.
Every element is, in a precise sense, the inversion of the corresponding element in a pitch competition.
The inoculation move
The cleanest way to explain what the Race actually does is to borrow a metaphor from medicine. The Constellation is a programme of investor-rejection inoculation.
Each milestone exposes the company, under controlled conditions, to one of the fatal-flaw tests that angels apply. The cap table milestone is the cap table inoculation. The IP assignment milestone is the IP inoculation. The SEIS Advance Assurance milestone is the tax-relief inoculation. The Q&A stress test is the handling-hostile-questions inoculation. A racer who completes all eight has been inoculated against the eight most common deal-killers documented in UK angel research.
This framing matters because it corrects something founders routinely misunderstand. Completing the Constellation does not make you investable. It makes you survivable. The frame is subtractive, not additive.
You are not becoming a better version of yourself; you are eliminating the specific weaknesses that would otherwise get you rejected.
What the angel sees (the constellation of milestones complete, the pattern lit up) is the signal that the inoculation has worked. What the founder does (the twelve months of eliminating rejection reasons one by one) is the mechanism. Both metaphors are accurate. Each does a different rhetorical job. The investor sees a constellation. The founder runs an inoculation.
This matters commercially because it gives both parties something they can evaluate. The angel is not trying to judge a founder in twelve minutes on whether they might be the next Brian Chesky. The angel is watching, over twelve months, whether the fatal-flaw checklist goes to zero. The founder is not trying to make themselves look extraordinary; they are trying to become legible. These are tractable problems.
The twelve-minute pitch was not.
The deeper point
Sutherland’s observation about design problems masquerading as other kinds of problems applies here with unusual force. The startup ecosystem has spent thirty years treating the equity gap as a capital problem and the rejection rate as a founder-quality problem.
Neither is right. The capital exists. The founders exist. What did not exist, until now, was a mechanism that let the information flow between them in a way that matched how the decision is actually made.
Every design choice in the Race is an argument about mechanism design.
- The twelve-month length is a mechanism choice; it is the shortest window in which an angel can form the kind of longitudinal judgement Mason’s research says they are actually making.
- The public leaderboard is a mechanism choice; it creates the continuous observation surface that replaces the episodic pitch.
- The Constellation is a mechanism choice; it translates Mason’s non-compensatory taxonomy of deal-killers into a set of public milestones.
- The Golden Buzzer is a mechanism choice; it converts the observation surface into investment action at the moment of angel conviction, rather than waiting for an artificially-scheduled decision event.
- The graduation architecture is a mechanism choice; it extracts positive-sum value from the fifty founders’ twelve months, rather than letting forty-nine of them drop into the “lost year” category that kills programme reputation.
None of this is about making the pitch better.
It is about retiring the pitch as the primary decision surface and replacing it with something that matches the evidence about how investors actually decide.
An angel reading this argument should, on Mason’s own evidence, recognise it as the thing they have been complaining about for thirty years: a mechanism that gives them enough time to see the founder clearly, enough structure to avoid false-negative screening, and enough social proof to justify conviction.
A founder reading it should recognise that they have been participating, all this time, in a system whose design works against them, and that the Startup Race is the first UK programme deliberately designed to work for them instead.
That is not a claim about being better at something the ecosystem already does. It is a claim about doing something the ecosystem has not yet tried to do.
The £100,000 Startup Race is the structural opposite of a pitch competition. Every element is designed for how angels actually decide, rather than for how angels say they decide. After thirty years of evidence, that distinction has started to matter.
James Shoemark is CEO of The Startup Race. Previous essay in this series: Revenue is the Brightest Star. But Investors Navigate by Constellations. The Race is open for applications for the 2026 cohort at thestartuprace.com.