How a Badge Becomes Credible
Venture Kick offers a useful way to see a quality signal being built rather than merely announced. This Swiss competition takes very early ventures through repeated judging, public recognition, feedback, and increasingly large awards. The question is not simply what its winners receive, but which parts of the institution make an early endorsement informative and which parts make the final comparison mainly about cash. Understanding that architecture is necessary before interpreting the competition's reported effects.
Venture Kick is a credibility pipeline
Venture Kick is designed for ventures near the beginning of their lives. During the study period, participating companies generally had not yet been incorporated or had only recently incorporated, had not received funding from professional investors, and were expected to have a link to a Swiss university or research center. The population is therefore narrower than “all startups.” It is a group of ventures seeking an early external assessment before the usual investor market has fully examined them.
That eligibility rule is the first component of the signal. A Venture Kick endorsement does not claim that an established company has defeated every competitor in its market. It says that a very young venture, meeting a particular institutional definition, has passed an evaluation at a moment when outside information is scarce. The meaning of the logo depends partly on this context. A reader who sees it knows something about the stage of the company and the kind of screening it has undergone, even before learning the full business story.
The second component is the judging pool. Venture Kick draws judges from investors, startup experts, and industry representatives. Most list expertise in science-based technologies, while a smaller share list internet, mobile, and software as their main area. This composition is not a footnote. It creates a potential fit between the institution's information-producing capacity and some technologies, and a potential mismatch with others. A panel is more likely to produce a useful signal when its members can interrogate the evidence that actually determines success in the venture's field.
The competition then repeats the evaluation across three stages. In the first stage, a startup presents its idea to a jury of roughly nine judges. Four winners receive a $10,000 award, public promotion, the right to display the Venture Kick logo, and access to the first two of the program's camps. The award is meaningful at a pre-incorporation stage, but the public endorsement and the first expert feedback are the distinctive additions. For a venture that has not yet attracted professional investors, the change from being invisible to being publicly recognized can be larger than the change in its cash balance.
Three months later, the second stage asks the surviving startups to present their business case to a different set of judges. Two are selected for another $20,000, further publicity, and access to the final camp. The new jury creates another assessment, but the publicity is now less novel. A company that has already been selected has an established relationship with the institution and a public record of progress. Stage 2 can reinforce the signal and provide learning, yet its marginal informational contribution is smaller than the first moment at which the market sees the venture endorsed.
Six months after that, two startups compete in the final stage. The winner receives the $100,000 final prize, bringing the potential total to $130,000, as well as another public announcement. This is a large change in cash relative to the earlier awards. It is a smaller change in visibility. Both finalists have already survived two rounds of scrutiny, appeared in Venture Kick publicity, and accumulated the credibility attached to earlier wins. The final announcement still matters, but its incremental certification is not comparable to the first public distinction between a recognized startup and one that did not advance.
The interface between these components is what makes the institution interesting. Judges generate an assessment; Venture Kick converts that assessment into a public result; founders receive feedback and resources; investors can observe the institution's endorsement; and the company moves through time with more or less runway. The badge is credible only if the judging process is trusted. The cash is useful only if its amount and timing fit the venture's costs. The stages do not operate as independent modules: changing one can change what the later stages mean.
The judging process gives the signal a more definite meaning than a popularity vote. In the first stage, each judge grades multiple aspects of the people and the project, along with a separate overall impression. Scores are recorded on a common scale and combined into a measure of startup quality for the research. The point is not that a numerical grade captures a venture's true value. It is that several judges, applying recurring criteria, create a more structured observation than an investor would get from a founder's claim alone. The time spent examining a small group of ventures also matters: a panel with enough attention to ask difficult questions can produce a different signal from a mass application screen.
The pipeline also gives participants information even when they do not advance. After each stage, Venture Kick provides successful and unsuccessful startups with private feedback that includes grades, comments, presentation material, and open questions about the project. This matters because a competition can allocate information as well as money. A losing team may learn that its market evidence is weak, that its technology is not yet convincing, or that its plan is too difficult to finance. Whether the team continues or stops, the institution has changed the information available to it.
The stages create a comparison—but not a clean split
The researchers use this architecture to estimate two effects. They treat winning stage 1 as a certification-dominated event because it is usually the startup's first public endorsement and comes with only a relatively small amount of funding. They treat winning stage 3 as a cash-dominated event because the finalists have already received earlier certification, publicity, and feedback, while the final prize is much larger. The logic is comparative: the first win changes visibility sharply, while the final win changes the cash position sharply.
This does not mean stage 1 is “badge only” or stage 3 is “cash only.” Even $10,000 can matter to a venture that has not yet incorporated. Stage-1 winners also receive camps, feedback, and promotion. The researchers therefore describe the estimate as certification-dominated, not as a pure certification treatment. Similarly, the final winner receives another endorsement and more publicity, even if those additions are subject to diminishing returns. The stage-3 estimate is predominantly about the marginal cash prize, but it retains a residual certification component.
The difference in visibility illustrates why the comparison is plausible. A stage-1 win changes whether the startup is legible in the wider startup information environment because it supplies the first public recognition and publicity. By the final stage, both the winner and the loser have already been publicized and have appeared in the competition's earlier process. The final win adds less new information about whether the institution has noticed the company at all. It mainly adds resources to two ventures that have already been screened.
To compare winners with appropriate counterparts, the study uses the detailed grades assigned by judges. These grades cover multiple criteria and are combined into a quality measure that predicts later outcomes. The underlying problem is selection: ventures that judges think are better are more likely to win, and better ventures may perform better even if they receive no reward. If the analysis simply compared winners with every loser, it would mix the effect of winning with the effect of having been judged stronger in the first place.
The grade is not a public ranking that investors can directly use. It is a research measure of what the judges observed before the award. In the study, a one-standard-deviation increase in the stage-1 grade predicts higher later survival, especially in the first several years. That predictive relationship gives the grade some value as a proxy for underlying quality, even though it is imperfect. The researchers also add year, competition, industry, and university controls, then check the main conclusions with alternative grade constructions, models that allow grading error, propensity-score matching, and a regression-discontinuity design around the stage-1 cutoff.
Using a proxy changes the meaning of the estimate. The analysis is not comparing identical startups, because quality cannot be observed directly; it is trying to compare ventures after accounting for the best structured measure available in the competition data. If grades contain noise, the estimates may lose precision. If the noise differs by technology, the comparison between science and ICT ventures becomes more delicate. The authors address this by allowing for imperfect grading and by using a cutoff design that relies less heavily on the grade itself, but no method turns the proxy into a perfect measure of quality.
The cutoff design is particularly intuitive. In stage 1, the startup just above the advancement threshold receives the treatment of moving forward, while the startup just below it does not. If those ventures are close enough in observed rank, a large jump in later outcomes at the cutoff is harder to explain by broad differences in quality alone. The study reports a discontinuity in survival and funding that is concentrated among science-based startups. This supports the idea that the early endorsement does something beyond identifying the ventures that were already strongest, although the small number of competitors per round makes the design less conventional than a large lottery.
The final stage cannot receive the same treatment. Only two startups compete for the last prize, and those two were selected from an already twice-screened group. The authors cannot implement a comparable final-stage regression discontinuity, so they treat the cash estimates as suggestive rather than definitive. Their confidence comes from the consistency of the pattern across grade-controlled regressions, errors-in-variables models, matching, and the timing of the observed effects—not from random assignment at the final stage.
That distinction between stages is easy to lose if the story is reduced to “Venture Kick winners did better.” A winner is both a recipient of a reward and a venture that passed a selection process. Stage 1 is informative because the marginal change is from low visibility to a first endorsement. Stage 3 is informative because the marginal change is from two highly screened finalists to one with much more cash. The design helps isolate the mechanisms, but it does not erase the bundle of publicity, feedback, networking, learning, and selection that accompanies participation.
Those shared activities create an important boundary. If winners and losers receive feedback or access to the same camps, those benefits may improve all participants without explaining the difference between them. That does not make the activities irrelevant; it means the stage comparison is focused on what winning adds at the margin. Likewise, a later-stage winner may receive a larger payment while the losing finalist receives no new public discovery benefit, but both have already crossed the institution's earlier filters. The comparison is strongest when read as a difference in marginal reward, not as a full account of everything the competition did.
The institution also limits how far the result can travel. The startups were unusually early, connected to Swiss academic institutions, and seeking a particular kind of support. The judges used a consistent framework, but their expertise was anonymized, so the data cannot show exactly which judge's knowledge produced which decision. A different panel might be better at assessing ICT ventures or less able to distinguish scientific quality. The Venture Kick badge is therefore not a universal property that can be detached from its evaluator, participant population, and moment in the startup lifecycle.
Those qualifications do not weaken the value of the design; they identify what it can teach. Venture Kick turns credibility into an observable institutional process: a defined population, a panel with a particular knowledge mix, repeated evaluations, public recognition, feedback, and rewards whose marginal meaning changes over time. The next question is what the resulting stage-1 and stage-3 comparisons actually show about survival, financing, technology, and persistence. Only then can the badge-versus-cash contrast be judged on outcomes rather than intuition.