THE THESIS
Why power-law dynamics demand a category of software that has never existed, and what changes when it does.
By Clint Browning, Founder of TheCipher
TheCipher exists to increase the percentage of investments that become fund-returning outcomes by compounding institutional judgment.
Every experienced investor understands the shape of returns in private markets. A small number of deals generate the vast majority of value, while the majority of investments return modestly or fail. This pattern is not an anomaly. It is consistent across asset classes.
| Asset Class | Typical Portfolio | % Investments Driving Most Returns |
|---|---|---|
| Venture Capital | 25–35 companies | 5–10% (1–2 carry the fund) |
| Growth Equity | 12–18 companies | 20–30% drive most economics |
| Buyout PE | 8–12 companies | Top 2–3 deals generate performance |
| Family Office | 10–20 concentrated positions | Top 20% transformational |
In venture capital, elite firms expect 50% or more of their investments to return less than cost. One company may return the entire fund. In buyout PE, the numbers are more forgiving, 70–90% of investments should create positive returns, but the top two or three deals still generate the majority of fund performance.
The math is uncomfortable, but most institutional investors have internalized it: somewhere between 80–95% of their portfolio is not going to define their fund's legacy.
Here is what I find remarkable: almost no one in the industry has built infrastructure designed specifically to increase that 5–10%.
The industry has built tremendous infrastructure for everything else. There is sophisticated tooling for deal sourcing, CRM, portfolio monitoring, LP reporting, and document management. There is AI for market intelligence and comparable transaction analysis. There are consultants, advisors, and operating partners at every stage of the investment lifecycle.
But almost no infrastructure exists to answer the only question that actually determines fund performance at the highest level:
Can we increase the percentage of our investments that become fund-returning outcomes?
The reason is subtle. Most institutional software is built for process efficiency. And process efficiency, while valuable, is not what separates a 3x fund from a 5x fund. What separates them is judgment, specifically, the compounding of judgment over time, across partners, across cycles, across asset classes.
Suppose a $1 billion fund makes 30 investments. In a typical outcome distribution:
| Outcome Type | # Investments | Return Multiple | Value Created |
|---|---|---|---|
| Failures | 15 | 0x | $0 |
| Average | 10 | 2x | $~667M |
| Good | 3 | 5x | $~500M |
| Great | 2 | 20x | $~667M |
Total portfolio value: approximately $4 billion. A 4x gross return.
Now suppose better institutional judgment changes three decisions:
Total portfolio value rises to approximately $5–6 billion. The fund may return 5x or 6x gross.
That is not a modest improvement. In a $1 billion fund carrying standard 20% carry, the difference in GP economics alone is hundreds of millions of dollars.
Small improvements in judgment produce outsized gains because investing is nonlinear. The return curve is not a straight line. It has a long tail. And every investment you move from the middle of the distribution toward that tail compounds dramatically.
SECTION 2
Section 1 established the math. Power-law returns are not a quirk of investing. They are the structure of it. A small number of decisions carry the fund. Now the harder question: where do those decisions actually come from?
Not from access. Every serious firm I've placed talent at has access: to deal flow, to data, to diligence resources that didn't exist a decade ago. Not from intelligence, either. The investment professionals I've recruited into PE and VC over 25 years are, without exception, sharp. Access and intelligence are table stakes. They explain why a firm gets into the room. They don't explain why one firm's outcome in that room compounds into a 20x and another firm's doesn't.
What explains it is judgment, and judgment is a specific thing, not a vague one. It is pattern recognition built from lived repetition: having seen a founder's behavior under pressure before, having watched a market dynamic unfold in a prior cycle, having sat in the room when a similar deal went sideways and knowing precisely which warning sign mattered and which one didn't.
Here is what's strange: every firm believes it has good judgment. Most are right, at the individual level. The partner who has been doing this for fifteen years genuinely does have pattern recognition that a generalist doesn't.
But judgment at the individual level and judgment at the institutional level are not the same asset. Individual judgment lives in one person's head, activates when that person happens to be in the room, and degrades the moment that person is distracted, on a different deal, or no longer at the firm.
Institutional judgment is different in kind, not just in scale. It is judgment that persists independent of who happens to be available on a Tuesday. It is the pattern from the 2019 deal showing up automatically when a 2026 deal rhymes with it, not because someone remembered to mention it, but because the system surfaced it.
Almost every firm has individual judgment. Almost no firm has institutional judgment.
In conversations with hundreds of investment professionals, the failure pattern is remarkably consistent. It shows up in three forms:
None of these are character flaws. They are structural consequences of judgment having no infrastructure. A firm can have brilliant partners and still lose the compounding effect of their judgment, for the same reason a company can have brilliant engineers and still lose institutional knowledge if nothing captures what they learn.
This isn't a new problem. What's new is that it's now solvable in a way it wasn't five years ago.
Capturing judgment used to mean asking partners to do more administrative work: log this insight, write up that pattern, tag this memo for future reference. That approach has always failed, for an obvious reason: the people whose judgment is most valuable are also the people with the least time and the least patience for additional process.
What changed is the ability to capture judgment ambiently, from the conversations, memos, and decisions that already happen, without asking anyone to change how they work, and to structure that captured judgment so it resurfaces automatically at the moment a new decision rhymes with an old one.
That capability is the foundation of everything that follows in this paper.
SECTION 3
If judgment is the asset that actually drives fund-returning outcomes, and most firms have plenty of it at the individual level, the natural next question is: why hasn't the industry already solved this?
I don't think it's for lack of trying. I think it's because the obvious solutions all fail in the same specific way.
A CRM tracks relationships and deal status. It answers "where is this deal in the pipeline" and "who do we know at this firm." It was never built to answer "does this situation resemble one we've seen before, and what happened when it did." Asking a CRM to do that is asking a filing cabinet to think.
A knowledge base is searchable, which sounds like progress, until you account for the fact that nobody searches for a pattern they don't know they're looking for. The value of institutional judgment is precisely in the moment a partner doesn't realize a current situation rhymes with a past one. A system that requires the user to already suspect a connection exists, in order to find it, isn't solving the recall problem. It's just organizing the haystack.
Modern document search, even AI-powered search, retrieves what you ask for. It is excellent at answering direct questions. It is not built to surface what you didn't think to ask. The most valuable interventions in an investment process are rarely the answer to a question someone posed. They're the warning nobody asked for, delivered at the exact moment it would have changed the outcome.
All three categories share the same blind spot: they are retrieval tools. They wait to be asked. Institutional judgment, when it works, doesn't wait. The partner who has seen this pattern before doesn't wait for someone to query them. They volunteer the insight unprompted, at the moment it matters, because they recognized the pattern in real time.
The tools the industry has built are retrieval systems. What's missing is a resurfacing system.
That distinction is not semantic. A retrieval system requires the user to know what to look for. A resurfacing system recognizes the pattern on the user's behalf and brings it forward, the way a sharp partner would, if that partner had perfect memory of every deal the firm has ever touched and was in the room for every conversation.
Building something that resurfaces judgment rather than merely retrieving it requires three capabilities working together, none of which alone is sufficient:
No single one of these is novel in isolation. Historical document ingestion exists. Ambient transcription exists. Pattern matching exists. What has not existed, until now, is the three of them built together, purpose-built for the specific shape of institutional investment judgment, not adapted from a generic productivity tool.
That is what I built TheCipher to be.
SECTION 4
TheCipher is built around a single mechanism: the Compounding Judgment Engine™. The name is literal, not a marketing flourish. It describes exactly what the system does: it takes the judgment a firm already has, in whatever form it currently exists, and ensures it compounds instead of evaporating.
Compounding, here, means the same thing it means in returns. A single insight, captured once, is worth something. The same insight, automatically resurfaced at every future decision point where it's relevant, across every partner, across every cycle, is worth something categorically larger, because its value doesn't decay with time or turnover. It accumulates.
The Engine operates on a continuous loop with three stages.
| Stage | What Happens | What It Solves |
|---|---|---|
| Capture | Historical documents are ingested; live conversations and meetings are transcribed ambiently | Judgment is no longer trapped in one person's memory or one unread memo |
| Structure | Captured material is organized into searchable, pattern-matchable institutional memory | Raw history becomes usable signal instead of an unsorted archive |
| Resurface | Relevant patterns are surfaced automatically at the moment a new decision resembles a past one | Judgment arrives unprompted, at the point where it can still change the outcome |
Each stage depends on the one before it. Capture without structure is just a larger archive, the same knowledge-base problem described in Section 3, at greater scale. Structure without resurfacing is a well-organized library nobody visits at the right moment. The loop only works as a loop. Remove any stage and the system reverts to a retrieval tool.
It's worth being precise here, because the market is full of AI tools claiming to serve private capital, and most of them solve a different problem than the one this paper has described.
Most AI tools built for PE and VC are productivity tools wearing an AI label. They summarize documents faster, draft emails faster, search faster. These are genuine improvements to workflow efficiency. They are not what this paper is about.
The Compounding Judgment Engine is not trying to make any single task faster. It is trying to change which investments become fund-returning outcomes, by ensuring the firm's collective judgment is present at the decision, not absent from it. That is a different category of claim, and it requires a different kind of system: one built around memory, pattern recognition, and timing, not around speed.
Efficiency tools make the firm faster. The Compounding Judgment Engine makes the firm righter, more often, on the decisions that matter most.
The Compounding Judgment Engine is not one feature. It is three components working together, each addressing one stage of the Capture → Structure → Resurface loop, plus a fourth that makes the output usable in the room where decisions actually get made:
Each of the next four sections takes one of these in turn.
This is why I believe TheCipher represents a fundamentally different category of institutional software.
Most software companies selling into PE and VC promise efficiency: faster workflows, better data access, reduced administrative burden. These are genuine improvements. They are also, ultimately, beside the point.
Institutional investors do not earn their returns by being more efficient. They earn their returns by being more right, more often, on the decisions that matter most.
TheCipher exists to increase the percentage of investments that become fund-returning outcomes by compounding institutional judgment.
That is a very different conversation than selling AI software. It is much closer to selling alpha.