The Useful Signal Is Not the Winner
This series examines a narrow but consequential question in early-stage venture screening: when a new technology-based firm asks for support, can the customer relationships described in its business plan tell a decision-maker more about five-year survival than a reader’s gut feeling? The practical judgment is deliberately modest. The evidence concerns a candidate signal for continued operation, not a crystal ball for growth, returns, or a famous success story.
That distinction matters because screening happens before the facts are comfortable. A public support body, accelerator, angel, or venture fund may face more applications than it can investigate deeply. At that point, a business plan, pitch deck, or one-page description becomes an early filter. It is useful because it is available; it is dangerous because it is written to persuade. A reader may notice a compelling market story, an impressive team, or a confident forecast and then convert that impression into a funding decision without being able to say exactly which evidence drove the judgment.
The study at the center of this series asks whether an observable feature of those plans—what it calls the maturity of transaction relations—can make that decision more consistent. The focal subjects are early-stage new technology-based firms, or NTBFs, and the relationship being examined is between their described dealings with customers and other parties, the screening judgment made from their plans, and whether the venture is still operating five years later. The companion evidence adds an essential warning: survival and growth are not the same prediction problem. A venture may remain alive without becoming large, and future sales can remain difficult to forecast even when the risk of early exit becomes somewhat easier to see.
Screening before the facts arrive
A venture evaluation is not simply a search for the most exciting idea. It is an allocation decision under uncertainty. Support providers must decide which applicants deserve scarce money, introductions, advice, or time, often when the ventures have little operating history and may not even have created a legal entity. The focal study describes business plans as a common first source of information in this setting, while also noting that formalized and objective ways to evaluate them are limited. In practice, individual heuristics and intuition remain influential.
Intuition is not automatically foolish. Experienced decision-makers can notice patterns that are hard to articulate, and a written plan cannot contain every relevant fact about a team or market. The difficulty is that an unexamined impression is hard to compare across applications, hard to audit after the outcome is known, and vulnerable to the persuasive design of the document itself. If two evaluators prefer different cues, a support system may be allocating resources through personal criteria while believing it is evaluating the same evidence.
An objective measure, in the study’s sense, does not mean a measure that is perfectly neutral or complete. It means a feature associated with venture outcomes that can be defined and coded rather than left entirely to a reader’s momentary impression. Transaction relations are proposed for that role because a venture’s relationships with customers, financiers, partners, and human resources can reveal how it is interacting with the environment. A statement that a team has identified a target market is not equivalent to a market test, which is not equivalent to repeated turnover. These are different levels of contact with the world outside the plan.
That contact is relevant to a resource-constrained venture. A technology can be impressive and still fail to find a paying use. A plan can describe a large market and still provide no evidence that anyone will exchange value with the venture. Conversely, early customer relationships may show that the founders are learning from the market, adjusting the product, and assembling resources around a real need. Those are plausible reasons to examine the signal. They are not yet reasons to treat it as a cause of survival.
Survival is not the same as winning
The first discipline of this subject is to define the outcome before interpreting the predictor. In the focal study, survival is a five-year continuity measure. It is a common minimum that can matter to public support providers, angels, and venture capitalists even though they may ultimately care about very different things: a sustainable livelihood, employment, rapid growth, an exit, or a financial return. Survival is therefore a useful floor for comparison across heterogeneous ventures. It is not a universal definition of success.
The distinction becomes clearer in the companion study, which follows 6,579 UK new ventures that began trading in the same quarter and tracks them for up to ten years. Its theoretical model treats sales growth as a sequence of shocks—close to a random walk—while treating survival as a question of whether the venture has enough resources to withstand those shocks. A small initial buffer can be exhausted quickly; a larger or accumulated buffer can allow a firm to remain in operation through bad periods. As the most fragile firms leave and survivors build resources, survival can become somewhat easier to classify.
That does not mean the surviving firms are on a predictable path to exceptional growth. In the UK data, the explanatory power of growth regressions declined over the years after entry in the baseline analysis, while the explanatory power of survival regressions generally increased. Among firms that survived the full ten years, the ability to explain growth remained roughly constant rather than improving. The authors’ conclusion is not that time reveals which venture will become a winner. It is that the “fog” can thin for the narrower question of continued operation while remaining thick around sales growth.
This is the boundary that keeps the customer-transaction argument honest. If a signal is associated with survival, it may help a screener identify ventures less likely to disappear early. It does not follow that the signal identifies the fastest-growing company, the best investment, or the next category leader. A venture can survive at a modest scale; another can grow rapidly for a period and then fail. A binary outcome and a winner’s trajectory should not be smuggled into the same word, such as “success.”
The diagram captures the central separation. A plan supplies an imperfect information surface. Customer transaction maturity is one candidate signal extracted from it, alongside human judgment and other context. The measurable question is whether that signal improves prediction of five-year survival. The arrow to growth or winner status is intentionally broken: the evidence in view does not establish that relationship.
The bounded claim this series will test
The focal paper’s claim is stronger than “numbers are good” and narrower than “algorithms choose winners.” It compares an objective, text-derived measure with subjective evaluation of business plans and with a deliberately agnostic benchmark. The objective measure is built from the maturity of transaction relations, with customer relations eventually providing the most useful incremental signal in the study’s model. The subjective comparison comes from non-experts reading plans, so it is a benchmark for intuition-driven assessment rather than a test of seasoned venture professionals.
The research therefore gives us several questions rather than one headline. What counts as a customer relationship at an early stage? How does the study convert a persuasive narrative into a comparable measure? What does five-year survival include, and what does it leave out? Does the signal outperform a simple prediction based on the sample’s overall survival rate? Does it outperform the particular intuition benchmark used in the study? And, even if it does, what kind of decision is the result strong enough to support?
Those questions will also keep the result connected to the venture itself. The point is not to replace the NTBF with an abstract score. It is to understand whether a specific feature of the venture’s relationship with customers reveals something that a plan reader might otherwise miss. The value of the signal, if it has one, lies in making a screening conversation more explicit: Which customers have actually interacted with the product? At what stage? Is there evidence of testing, repeat exchange, or turnover, or only a description of an intended market? What should be verified before scarce support is committed?
The answer will not be a universal cutoff. The focal sample contains 272 technology-based ventures from one German region and one high-tech business-plan competition. Its comparison with non-experts is informative, but it does not tell us how experienced investors would perform. The source also does not show that using the measure in a live organization improves funding outcomes. These limits are not decorative caveats; they determine how far the evidence can travel.
The useful conclusion, then, has two parts. Customer transaction maturity may be a more disciplined early survival signal than raw gut feel in the study’s bounded comparison. But a survival signal is a floor, not a forecast of greatness. To see why both statements can be true, the next step is to examine the different ways survival and growth become visible over time—and why the distinction matters before any customer relationship is turned into a score.