This article is seventh in a Deal Screening Mastery series for emerging investors who want to build LP confidence and for early‑stage companies trying to find the right investor fit. In the sixth article, “Follow the Path, Not the Pitch,” I showed you how to assess regulatory and reimbursement readiness at every funding stage.
A strong medical product pitch can be full of motion without containing much proof. Founders can show pilots, marketing collateral, engagement charts, early study results, and even early signs of revenue, yet still leave the most important investor question unanswered: what, exactly, should this evidence cause a serious underwriter of risk to believe?
Investors do not get paid for admiring activity. They get paid for upgrading conviction accurately. Once a company has a plausible timing story and a believable regulatory or reimbursement path, the next screening question is whether reality is starting to line up with the narrative. This is where many healthcare deals become deceptively hard to read. The data may not be fabricated. The founders may be sincere. The market may even be large. But the evidence package is often misaligned with the claim it is trying to support. A company may present engagement metrics when the real gating issue is payer economics. Another may highlight promising preclinical work while speaking as though clinical utility is already established. A third may point to early revenue that looks encouraging until it becomes clear those dollars came from custom services, one-time implementation work, or a champion-driven relationship that is unlikely to repeat.
The discipline, then, is not to ask whether something is happening. It is to ask what kind of belief the available evidence rationally supports. In medicines and medical devices, that question is especially important because no single metric means much outside the context of regulatory path, reimbursement dynamics, buying process, workflow adoption, and clinical relevance. Evidence is never generic. It is always tied to a specific risk and stakeholder.
The cleanest way to think about evidence is as a mechanism for updating belief. Every relevant datapoint should reduce uncertainty around a specific category of risk: technical feasibility, clinical validity, safety, regulatory viability, reimbursement potential, operational adoption, or commercial repeatability. If the evidence does not clearly change one of those beliefs, it may still be interesting, but it is not yet underwriting conviction. That framing helps investors avoid one of the most common screening mistakes: evaluating evidence in isolation instead of evaluating evidence against the claim being made. No evidence is inherently good, validating, or commercially relevant. That pilot, statistically significant result, or purchase order becomes meaningful only when linked to a concrete question such as:
Does this prove the product works in the intended setting?
Does this show the buyer will pay repeatedly?
Does this indicate the implementation burden is manageable?
Does this suggest a payer has a reason to care?
Different healthcare businesses require different evidence packages. If the path runs through payer coverage, the most important evidence may center on outcomes, utilization, total cost of care, or resource use, rather than app engagement or physician enthusiasm. If the path depends on provider adoption, workflow fit and behavior change may matter as much as efficacy, because a clinically sound tool that does not integrate into care delivery can still stall commercially. If the company is building a diagnostic reimbursement story, analytic validity and clinical utility should outweigh generalized excitement about the novelty of the technology. And if the business touches multiple constituencies, there must be evidence that maps to each of them, rather than assuming one favorable signal transfers automatically to the rest.
From a founder’s viewpoint, three questions can sharpen almost every evidence discussion:
What belief is this data trying to change?
How strong would the data need to be to change it?
Who, in the real world, truly cares about this metric?
Those questions sound simple, but they expose a surprising amount of weakness. Founders who have internalized them tend to present evidence with precision. Founders who have not often default to collage: many slides, many data points, many activities, but limited clarity on what any of it proves.
One reason investors over- or underreact to evidence is that they judge it against the wrong development stage. A preclinical company should not be expected to have mature commercial proof. A commercial-stage company should not be excused for fuzzy customer evidence simply because healthcare sales are hard. The standard must adjust with the maturity of the business.
At the earliest stage, “good” evidence usually looks like crisp technical milestones, relevant model systems, reproducible bench or animal data where applicable, and early safety or tolerability signals if those are part of the core risk. The point is not to pretend the company has de-risked the downstream clinical or commercial story. The point is to show that the foundational technical claim is becoming more credible in a disciplined way. This is also where overclaiming becomes a serious tell. A company that only has early lab work but speaks as though clinical utility, physician adoption, or payer ROI are already implied may be collapsing multiple future proof points into one present-tense narrative. That creates a screening problem because it hides where the real uncertainty still lives. A useful investor question at this stage is whether the evidence lines up with the stated maturity of the company. If the answer is yes, the business may be appropriately early rather than underdeveloped. If the answer is no, the team may be using technical promise to borrow credibility from milestones it has not yet reached.
In the early clinical or pilot phase, perfection is not the standard. Small studies, single-site work, non-randomized designs, and narrow populations are often normal. What matters more is whether the work is coherent. Is the population defined? Is there a comparator, even if imperfect? Are endpoints understandable and relevant? Can management explain what the study can prove, what it cannot prove, and what would need to happen next? Strong early-stage teams understand the difference between encouraging data and definitive data. They do not try to hide sample size limitations, selective recruitment, or context-specific findings. Instead, they show that the signal is directionally useful and that the next experiment, study, or contract is designed to answer the next uncertainty. Weak teams often make the opposite move. They convert every positive directional outcome into the language of inevitability. They speak as if a pilot with ten motivated users predicts broad deployment, or as if a promising subgroup result can stand in for generalizable impact. Investors should not punish a company for being early. They should, however, discount companies that cannot articulate the boundaries of what their own data means.
At commercial stage, the evidence burden changes again. Here the key issue is not simply whether someone paid. It is whether the payment and usage pattern resemble the long-term business model. A clear paying customer, a defined use case, repeat utilization, renewal behavior, implementation progress, or expansion within an account can all be meaningful. Messy contracts, early discounts, and imperfect unit economics are common and not automatically disqualifying. The correct question is whether early dollars align with the future business. If the company says it is building a scalable software or platform business, but most revenue comes from bespoke services, custom analytics, implementation-heavy work, or founder-dependent selling, then the traction may be less transferable than it appears. That is why early revenue can be either one of the strongest or most misleading signals in a healthcare deck. Revenue is emotionally persuasive. It feels like proof because money changed hands. But unless it reveals something durable and repeatable about buyer behavior, economic ownership, integration, and renewal potential, it can still be mostly decorative, even if it extends company runway.
Healthcare founders and investors are both susceptible to vanity metrics. Certain datapoints create a fast impression of progress, and once that impression forms, it becomes difficult to resist building a bullish story around it. The antidote is to understand the most common false positives before they become embedded in the mental model. Let’s review them next.
Digital health decks frequently emphasize logins, downloads, messages sent, time spent, and feature usage. Those numbers can be useful operational indicators, and are easily measured, but on their own they are rarely investable evidence. Engagement is only meaningful if it connects to the economic or clinical outcome that the real buyer values. For example, high patient interaction may matter if it leads to measurable adherence, reduced avoidable utilization, lower administrative burden, or improved performance on a reimbursed metric. If it does not, engagement may simply mean the product is active, not that it is valuable. Investors should ask what downstream behavior or outcome the engagement signal is supposed to predict and whether that relationship has been demonstrated.
Many pilots can look impressive in healthcare because pilots are hard to get and partner logos confer social proof. But pilot count alone is one of the weakest screens available. A pilot with no budget owner, no predefined success criteria, no implementation commitment, and no path to conversion is often best understood as market exploration, not validation. The more useful questions are operational.
Who owns the budget if this works?
What metric determines success?
How many pilots converted?
How many expanded?
How many died quietly after the “evaluation phase”?
A company with fewer pilots but cleaner conversion logic may be far stronger than one with many loosely structured experiments.
Clinical evidence can also create false confidence when the comparator is missing, weak, or inappropriately selected. Pre/post designs in volatile environments, best-site analyses presented as if they are representative, and subgroup slices elevated above the total dataset are all common ways the story becomes stronger than the evidence. This does not mean all early clinical evidence is suspect. It means investors should stay disciplined about context. Compared with what? In whom? Over what period? Using which endpoint? Under what operating conditions? If it’s hard to understand what was compared, in whom, under which conditions, in what time frame, and with which endpoints, the data may be more promotional than informative.
Revenue deserves special skepticism because it can conceal as much as it reveals. One-time implementation fees, grant-like partnership payments, sponsored pilots, and custom projects can all show up as top-line traction. Yet those dollars may say little about whether a scalable commercial engine exists. A simple test is to ask what must happen for the next ten customers to look meaningfully similar to the last three. If the answer depends on founder heroics, unusual relationships, custom scope, or one-off strategic exceptions, the revenue may be real, but the business model is still immature. Real traction is not just paid activity. It is paid activity with repeatable logic.
The best evidence slides make it easy to understand the question, the design, the result, and the limitation in one clean pass. They make a deck feel evidence-ready rather than evidence-adjacent. Such a deck typically does four things well.
It states the question clearly. What hypothesis was being tested or what operational claim was being evaluated?
It explains the design. Who was studied, in what setting, against what comparator, over what timeframe, and with which endpoint?
It presents the result with context. The reader should know not just that something improved, but relative to what baseline, standard of care, operational benchmark, or published norm.
It names the limitations. Small sample, short follow-up, single-site context, selection bias, or implementation constraints should be surfaced directly rather than buried in fine print.
This is where basic statistical hygiene matters, regardless of stage. Investors need enough structure to distinguish signal from noise. Founders often worry that naming limitations will weaken the story. However, honest framing increases credibility because it shows the team understands the burden of proof and knows what still must be earned. In a space as complex as healthcare, intellectual honesty is itself a positive signal.
Traction becomes meaningful when it reflects repeatable behavior from the stakeholder who matters most in the company’s path to scale. That is why investors should look beyond surface indicators and ask whether the observed activity demonstrates durable value transfer. Some traction signals are materially stronger than others. Renewals indicate value was experienced over time rather than merely anticipated up front. Expansions suggest the initial use case led to broader organizational confidence. Multi-site rollouts show an organization was willing to operationalize, not just experiment. Movement toward the economic buyer suggests the product is climbing from local enthusiasm toward budget authority.
The definition of strong traction also changes with the company’s commercial path. For payer-driven models, good traction may involve pilots tied to outcomes, utilization, claims-based performance, or total-cost-of-care hypotheses. For provider-driven models, workflow adoption and implementation behavior may matter more than broad top-of-funnel interest. For employer or direct-to-consumer models, persistence, retention, cohort behavior, and downstream conversion often matter more than acquisition spikes. This is why partner logos deserve interrogation rather than admiration. A recognizable health system, payer, employer, or pharma name can be a meaningful signal, but only if the relationship maps to the company’s actual ideal customer profile and future sales motion. Investors should ask about the contract type, contract size, implementation scope, use case, buyer level, and renewal history with the partner. Otherwise, you risk brand halo substituting for commercial foothold.
One of the most important investor habits is refusing to treat evidence as universally transferable. The same study, contract, or usage pattern can mean very different things depending on the company’s path through the market. A clinically enthusiastic user base does not solve a payer reimbursement problem. Strong app engagement does not answer a provider workflow challenge. Positive pilot outcomes in a highly resourced academic center may not translate to a lower-resource community setting where the commercial rollout is planned. Trial endpoints that excite clinicians may still leave hospital finance teams or payers unmoved if they do not connect to budgetary or reimbursement-relevant metrics. This misalignment shows up constantly in company fundraising. Founders are often presenting true positives from the wrong frame. The data is not useless; it is simply not the evidence the next gatekeeper needs. Investors who recognize this early can avoid confusing “promising” with “sufficient.” That discipline also creates a more nuanced screening outcome. Sometimes the right conclusion is that the science looks better than the commercial story. Sometimes the commercial pull looks stronger than the clinical package. Sometimes the reimbursement logic is compelling, but the implementation evidence is still thin. Writing that distinction into screening notes is valuable because it prevents the overall narrative from becoming smoother than reality.
A practical first-pass screen does not need to be elaborate. Three questions can surface a large share of the signal:
What is the single strongest piece of evidence the company has, and what specific belief does it change?
What is the most oversold or potentially misleading datapoint in the deck?
If all activity stopped today, what repeatable behavior or value has actually been demonstrated?
Those questions help sort motion from proof very quickly. Deals with abundant activity but little belief-changing evidence should usually move down the priority list, even if they look busy.
Conversely, deals with modest activity but unusually strong and well-framed evidence may deserve more attention than their surface momentum suggests. This lens also works as a coaching tool for founders. The goal is not to shame teams for being early or imperfect. It is to help them understand what evidence the market will need next, what claims they have earned versus borrowed, and where decorative traction is crowding out the more demanding proof points that build investor confidence. In the broader deal screening process, this is the lens that asks whether reality is beginning to support the story. “Why now?” may tell you whether macro conditions are favorable. Regulatory and reimbursement readiness may tell you whether the path exists and whether the route is navigable. Evidence, signals, and traction quality tell you whether the company has started to earn belief rather than merely request it. That shift, from presentation to proof, is often where serious conviction begins.
The next article in this series will discuss market structure and Go-To-Market complexity. Don’t miss it!
Thanks for reading Thinking Kat! If you found this issue valuable, please pass it to someone else who would benefit.
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.