
I built Pelican Alpha after deciding to codify what was working in my own investing. I've been investing in public equities since 2018, and over the last 3 years I've nearly 8x'd my money investing primarily in micro and small caps.

Pelican exists to solve the single biggest bottleneck I faced as an investor: turning over enough stones to find real opportunities, then underwriting them before alpha decay sets in. Below is my background, the exact process I follow, and how I turned it into a system.

---

### My Background

I started my investing career at the multi-family office ICONIQ Capital, moved to the activist hedge fund Osmium Partners, and later managed a concentrated long-only strategy. Throughout my career, I've worked across activism, secular growth, long/short, and deep value strategies. My personal sweet spot — and where I've generated the most alpha — is in GARP (Growth at a Reasonable Price) and Special Situations.

I am a CFA charterholder with a background in economics and accounting from UCSB. I began my career in public accounting, which gave me the foundation I use today to analyze financial statements.

My strategies are inspired by Buffett, Munger, Greenblatt, Marks, Li Lu, Mohnish, Klarman.

---

### Track Record

Over the last 3 years, since leaving my last job, I've nearly 8x'd my capital investing primarily in micro and small caps. My Roth IRA is up 124% over the last 12 months vs the S&P at 21%. A second account I manage is up 44% over the same period. The two years before that, I nearly 4x'd my capital.

![Track record (2022–2024): quadrupled account value](/blog/track-record-2022-2024.png)

I also run a **public mock portfolio** on Savvy Trader where every position, entry, and exit is fully transparent: [savvytrader.com/ricardoleee93/pelican](https://savvytrader.com/ricardoleee93/pelican). The account is up 82% over the last year and 38% YTD (72% CAGR since inception in mid-2025). Follow it if you want to track my picks in real-time.

Some of my biggest winners in the small/micro space: ALTO, IMPP, TEN, IE, and GOOGL.

Past performance is not indicative of future results, and investing always involves risk.

---

### How I Think About Markets

I don't believe markets are fully efficient. As Joel Greenblatt has pointed out, businesses don't fluctuate in value by 50% or 80% in a single year, yet their stock prices often do. If the market were a precise weighing machine, these swings wouldn't exist.

Small cap investing is largely a game of turning over stones. The inefficiency is real — most institutional investors can't own these names because of liquidity constraints, which means they're under-researched and mispriced far more often than large caps.

Everything I do is valuation-based. As Mauboussin said, everything should be a DCF. Unless a company has serious bankruptcy risk, there is almost always a price where that business becomes a bargain. The question is always: what is the business worth, and can I buy it for meaningfully less than that?

My mental model is simple:

1. Look at a massive number of ideas (High Velocity).
2. Kill the bad ones quickly (High Filter).
3. Concentrate capital when you find the obvious winners (High Conviction).

---

### The Process: 5 Steps

---

### Step 1: Find Ideas Worth Investigating

You need a systematic process to surface ideas that are likely to be on the winning side of the trade BEFORE you even look at the fundamentals. Fish where the winners are.

What I screen for:

- **Insider trading clusters** — a CEO buying $5,000 is noise. A CEO, CFO, and chairman buying a combined $2,000,000 on the same day is a signal. Cluster buys are far more predictive than individual purchases. When the people who know the business best are putting their own money in, that's worth investigating.
- **Congressional trades** — love them or hate them, members of Congress trade on information advantages. These are public disclosures.
- **13F filings** — what are the funds you admire actually buying? Osmium, Greenlight, Baupost, Pershing Square — their quarterly holdings tell you where smart money is allocating. If Klarman is buying something small and illiquid, that's a strong signal. (I wrote a detailed breakdown of [how to read 13F filings](/blog/how-to-read-13f-filings) if you want to go deeper on this.)
- **Price target increases** — when an analyst raises a target from $5 to $16, that's worth investigating, even if you don't trust sell-side targets blindly.
- **Activist involvement** — when an activist takes a 5%+ position and files a 13D, there's usually a catalyst coming. This is classic special situations investing.
- **Material press releases** — new contracts, management changes, strategic reviews, divestitures.
- **Quantitative screens** — businesses with 3-year revenue CAGR north of 10%, ROIC north of 10%, trading under 15x EBITDA.

Every day I screen thousands of data points across these 7 categories to find the most interesting signals. Each gets rated on a scale of 1-10 based on the strength of the signal.

---

### Step 2: Gather Data and Understand the Business

Once you have names worth investigating, you need to do the actual work. This is the part most people skip or do superficially. You need to go through:

- **Earnings call transcripts** — what is management saying vs what the numbers actually show? Are they dodging questions? Are they sandbagging guidance? The transcript tells you things the 10-K never will.
- **Annual reports and 10-Ks** — business segments, risk factors, strategy, competitive positioning
- **Proxy statements** — how is management compensated? What are they incentivized to do? Do they own stock? Is their comp tied to revenue growth (potentially value-destructive) or ROIC (shareholder-aligned)? Munger said "show me the incentive and I'll show you the outcome" — the proxy is where you find the incentive.
- **Financial statements** — income statement, balance sheet, cash flow. Key ratios: ROIC, FCF margins, debt-to-equity, working capital trends
- **Competitor data and industry reports** — who are the peers, how does the target compare on margins, growth, valuation?
- **Press releases and 8-K filings** — material events, acquisitions, divestitures, management changes
- **Institutional holdings and insider transactions** — who's buying, who's selling, what's short interest?
- **Web search** — customer reviews, product sentiment, Glassdoor, industry news

For a single stock, you're looking at hundreds of pages of filings. This is where having a systematic process saves enormous amounts of time.

---

### Step 3: Value the Business

What you do with this data depends on the type of business.

If a company is trading below tangible book, you want to know what the assets are actually worth — so you perform a **liquidation value analysis**. This is Graham-style net-net investing. If a company trades at 20x book, liquidation value is completely irrelevant.

For most businesses, a **DCF** is the anchor valuation. You're projecting free cash flows, applying a discount rate, and estimating a terminal value. The key is getting the assumptions right — and being honest about what you don't know.

**Peer multiples** — what are comparable companies trading at? If similar businesses trade at 12-15x EBITDA, is there a reason your target should trade at 8x? Maybe the discount is deserved. Maybe it isn't.

**Historical multiples** — what has this business traded at over the last 5 years? If it normally trades at 18x earnings and it's currently at 12x, is that a signal or a value trap?

**Analyst price targets** — I don't trust sell-side targets blindly, but the consensus gives you a market check.

You weight each method based on how relevant it is to the specific business, and you build a mosaic of what the business is worth. Not all methods deserve equal weight — and different business types call for different approaches.

---

### Step 4: Adjust for Qualitative Factors

A raw valuation number isn't the final answer. You need to adjust for things the numbers alone can't tell you.

**Management quality and alignment** — does the management team own meaningful stock? Are their incentives aligned with shareholders? If they're compensated to grow revenue at all costs, they'll grow revenue at all costs — even if it destroys value. What's their track record at prior companies? (I wrote a detailed breakdown of [how I grade management teams](/blog/capital-allocation-scoring) using an archetype-first framework.)

**Moat durability** — does the business have pricing power? Are customers locked in? Is there a network effect? Is the moat widening, stable, or narrowing? Are they gaining or losing market share?

**Sentiment check** — who owns the stock? What's short interest? What are insiders doing — buying or selling? What are institutions doing — accumulating or distributing? What does the sell side think?

These qualitative factors should move your valuation up or down. If peers trade at 20x EBITDA but your target has misaligned management, a shrinking moat, and heavy insider selling — maybe it deserves 15x. There's a deserved discount, and your job is to figure out how much.

---

### Step 5: Make the Decision

After adjusting all your valuation methods, you arrive at a final implied value. Then, as Buffett said, if there's a sufficient margin of safety, you buy the stock.

If you make enough independent bets like this over time, you are likely to do well. The key words are "independent" and "over time." Concentrate when conviction is high, diversify across ideas, and be patient. As Li Lu has said, the ability to do nothing is one of the most valuable skills an investor can have.

---

### How Pelican Systematizes This

I found that doing this process manually was taking me 20-30 hours per stock. So I built Pelican Alpha to automate the entire workflow and give retail investors institutional-level equity research.

**1. Daily Signal Screening** — Pelican screens 7 signal categories daily across thousands of data points (insider trades, congressional trades, 13Fs, price targets, activism, press releases, quant screens). Each signal is rated 1-10 based on strength. For example, a cluster buy from the CEO, CFO, and chairman totaling $2M scores much higher than a single director buying $5K.

**2. Comprehensive Research** — For the highest-rated signals, Pelican runs 35+ analysis prompts using all the raw data I mentioned above: transcripts, annual reports, proxies, financial statements, competitor data, industry reports, and web search. It ingests filings into a vector database and uses RAG to analyze them.

**3. Scoring and Valuation** — It scores the business's moat, scores the management team, scores sentiment. Then it performs valuation using DCF, peer comps, historical multiples, and price targets — many adjusted for the qualitative factors above. Then it calculates a weighted average to estimate fair value.

**4. Context-Aware Analysis** — Pelican doesn't evaluate all businesses the same way. A biotech should be analyzed differently than an energy cyclical. For SaaS, it looks at ARR, retention, and rule of 40. For banks, it's net interest margin and loan-to-deposit ratios. A company experiencing rapid growth like BROS should be evaluated differently than a late-stage mature company like SBUX. So Pelican classifies every business into one of 5 archetypes: Early-Stage Hypergrowth, Mature Compounder, Cyclical Operator, Asset-Based Operator, and Turnaround Candidate. The archetype changes the entire analytical framework.

**5. AI Research Agent** — Subscribers also get access to an AI agent with 31+ data tools that can pull live financials, search SEC filings, run structured analysis frameworks, and answer equity research questions in real-time.

450+ hedge funds currently receive our daily signals every morning via email, and that list is growing fast. I decided to open Pelican up to retail investors because I think individual investors deserve the same quality of research that institutions get. The whole point is to level the playing field.

---

### A Personal Note

On a personal level, I was born in Brazil, I am ethnically Korean, and I moved to the U.S. at age 7. I'm currently living in Orange County with my parents and our dog, Xuxa.

When I'm not studying businesses, I'm usually playing basketball, watching MMA, or with my family.

I hope you enjoy the research.

Ricardo Lee
Founder, Pelican Alpha

---

### Contact

- [LinkedIn](https://www.linkedin.com/in/ricardo-lee-cfa-89a13812a/)
- [ricardo.lee@pelicanalpha.com](mailto:ricardo.lee@pelicanalpha.com)


Founder, Pelican Alpha

I 8x'd my portfolio in 3 years investing in micro/small caps. Here's the exact process I use and how I turned it into a system that 450+ hedge funds use daily.

Why I Built Pelican


Apple spent $437 billion buying back its own stock over the last five years. I scored that a 4 out of 4 — exceptional. ADM spent $6.5 billion buying back its own stock over the same period. I scored that a 2 — adequate. Both companies were shrinking their share count. Both were returning capital to shareholders. But one was doing it brilliantly and the other was doing it at the wrong time in the cycle.

The difference isn't the *action*. It's the **context**. And that's why most management scorecards are broken — they treat buybacks, dividends, and reinvestment as universally good or bad, when the reality is that what constitutes great **capital allocation** depends entirely on what kind of business you're running.

I built a management grading system into Pelican's research pipeline that classifies every company into a business archetype *before* scoring the team. The classification changes the rubric. Here's how it works, using three real companies I've analyzed.

---

### The Framework: Classify First, Then Score

Every company gets assigned one of five business archetypes based on its financials, growth profile, balance sheet, and management narrative:

- **Mature Compounder** — durable moat, self-funded, high ROIC, moderate growth (5-15%)
- **Cyclical Operator** — performance tied to macro or commodity cycles, high margin volatility
- **Asset-Based Operator** — value from physical/regulated assets, capital-intensive, stable cash flows
- **Early-Stage Hypergrowth** — sacrificing profitability to capture a large market, 20%+ revenue growth
- **Turnaround Candidate** — business in distress requiring restructuring to survive

Then I score management across four pillars — Capital Allocation Skill, Incentive Alignment, External Track Record, and Board Oversight — each on a 0-4 scale. The criteria for what earns a 4 versus a 0 shift completely by archetype.

Let me show you what that looks like in practice.

---

### Apple (AAPL): The Compounder Doing It Right — Score: 4/4

Apple is classified as a **Mature Compounder** — 45.7% ROIC, net leverage of just 0.3x, and $582 billion in operating cash flow over the last five years. The archetype test for capital allocation is: *are you compounding per-share value through disciplined buybacks and reinvestment?*

Apple is the textbook case. Of every dollar of capital deployed, 71.5% went to buybacks. The result: share count down 12% over 4.8 years, driving EPS growth of 76.7% (12.7% CAGR) — of which roughly one-quarter came from the shrinking share count and three-quarters from profit growth. That's the right mix. The buybacks are amplifying real operating improvement, not masking a declining business.

But the score isn't just about *how much* they bought back. It's about *when*.

Apple's buyback timing scores tell the story. In late 2022, when the stock was under pressure and EV/EBITDA compressed to ~19-20x (87th-89th percentile of cheapness versus its own 5-year history), Apple bought $44 billion of stock across two quarters. Those tranches have delivered 61-77% trade ROI. When the stock was more expensive in late 2024 — 25-28x EBITDA (5th-18th percentile) — they kept buying, but those tranches are up only 7-13%. They didn't stop, but they leaned in harder when it was cheap. That's discipline.

**Why this earns a 4 for a Compounder:** Tim Cook's 15-year tenure has produced $437 billion in buybacks at a 17.9% net yield, zero equity issuance, and M&A spending of just $330 million (0.1% of capital deployed) with 160.9x effectiveness. The board includes four former public-company CEOs on its committees. LTI compensation is tied to 3-year relative TSR versus the S&P 500 with a cap at 100% payout if absolute TSR is negative — meaning executives can't get paid for beating an index while the stock declines.

**Red flags I'm watching:** Insider selling of $130 million with zero open-market purchases in the last 12 months. And incremental ROIC turned negative (-140%) as AI infrastructure capex ramps — a signal worth monitoring if it persists.

---

### ADM: The Cyclical That Bought at the Wrong Time — Score: 2/4

Archer-Daniels-Midland is classified as a **Cyclical Operator** — ag commodities, high margin volatility, performance tied to crush spreads and commodity prices. The archetype test here is fundamentally different from Apple's: *did management allocate capital counter-cyclically — investing at troughs and returning cash at peaks?*

ADM fails this test. Look at the buyback timeline. In Q3 2022, with crush margins peaking and the stock at $82-92, they spent $1.2 billion buying back shares. Those tranches are down 18-27%. Then as the cycle turned and the stock fell into the $58-61 range by early 2024 — the cheapest it had been in years, with EV/EBITDA timing scores in the 94th-97th percentile — they bought another $2.3 billion. Better timing on those tranches (up 10-15%), but by FY2025, when the stock was at its cheapest and the balance sheet had $5.5 billion in operating cash flow from an inventory liquidation, they stopped buying entirely.

That's **pro-cyclical** capital allocation — the natural human instinct and the exact wrong behavior for a cyclical business. You want management that buys aggressively at the trough when cash is tight and sentiment is terrible, and conserves capital at the peak when everyone feels rich. ADM did the opposite.

The numbers confirm it. Net income declined 48% over the period. ROIC sits at 3.0% against an estimated 8% WACC — meaning the business is currently destroying economic value on its invested capital. M&A effectiveness is -0.4x, indicating that $2.5 billion in acquisitions *reduced* EBITDA rather than growing it.

**Why this earns a 2 for a Cyclical:** The deleveraging is real — net leverage down to 2.86x with $1 billion in debt repaid — and that's the right move for a cyclical in a trough. But the buyback timing, sub-WACC returns, and negative M&A effectiveness prevent a higher score. The board has strong commodity expertise (Patrick Moore in commodity management, Terrell Crews in agri-business operations), which earns a 3 on governance. But the incentive structure is a problem: specific STI and LTI metrics are undisclosed, so I can't confirm whether management is compensated on through-cycle ROIC or single-year peak earnings. For a cyclical, that distinction is critical — incentives tied to peak-year performance encourage exactly the pro-cyclical behavior ADM demonstrated.

---

### National Fuel Gas (NFG): The Asset Operator Playing Its Role — Score: 3/4

NFG is classified as an **Asset-Based Operator** — integrated natural gas across upstream, pipeline, and utility segments in the Appalachian Basin. Capital-intensive (70% of capital deployed to capex), regulated cash flows, investment-grade credit. The archetype test: *is management generating positive returns on newly invested capital while maintaining a fortress balance sheet?*

NFG passes. Incremental ROIC is 576.9% — meaning every new dollar invested is generating outsized returns. Net leverage is a conservative 1.7x with 6.4x interest coverage, investment-grade ratings from both S&P (BBB-) and Moody's (Baa3), and a dividend that has been raised for 55 consecutive years.

Buyback activity is modest ($144 million, just 2.2% of capital deployed) but impressively timed. The heaviest purchases came in Q4 2023 and Q4 2024 when the stock traded at $51-61 — EV/EBITDA timing scores of 83-94/100, meaning management was buying near the cheapest valuations in the company's 5-year history. Those tranches have delivered 40-66% trade ROI. When the stock was more expensive (Q3 2025, around $87), they pulled back to a token $524K purchase. That's opportunistic.

**Why this earns a 3, not a 4:** For an Asset-Based Operator, a 4 requires demonstrated returns on *every* new asset investment and deeply aligned incentive structures. NFG's overall ROIC of 6.3% is solid for a regulated utility but not exceptional. More importantly, insider ownership is just 1.32%, with most holdings in 401(k) plans rather than direct stock. No open-market insider purchases in the last 12 months. And the specific compensation metrics — the STI and LTI targets that would tell me whether management is incentivized on asset turnover or ROA versus generic revenue growth — aren't disclosed. For a capital-intensive business reinvesting 82% of operating cash flow back into assets, knowing *what management gets paid for* is essential. The lack of transparency caps the incentive alignment pillar at a 2.

The governance pillar, however, earns a 3: Barbara Baumann brings financial expertise and deep oil and gas experience, Jeffrey Shaw is a CPA and former utility CEO serving as Lead Independent Director, and 10 of 11 board members are independent.

---

### What This Framework Changes

If you evaluated AAPL, ADM, and NFG on a single generic checklist — "Are they buying back stock? Yes. Are they paying dividends? Yes. Is leverage manageable? Yes." — all three would score similarly. They'd all look "fine."

But they're not all fine. Apple is compounding per-share value with exceptional timing and a world-class board. ADM is buying back stock pro-cyclically while generating sub-WACC returns and hiding its compensation structure. NFG is doing exactly what an asset operator should do — reinvesting at high incremental returns and maintaining a fortress balance sheet — but has an alignment gap between management's modest ownership and the capital they're deploying.

Same actions. Different archetypes. Different grades. That's the point.

The next time you evaluate a management team, start by asking: *what type of business is this?* A CEO who initiates buybacks at a turnaround with 5x leverage isn't disciplined — they're reckless. A CEO who halts all buybacks at a hypergrowth company to reinvest every dollar in scaling isn't cheap — they're doing their job. The framework only works when the lens matches the business.

Every Pelican research report classifies the company into one of five archetypes and scores management with archetype-calibrated rubrics. If you want to see this applied to a specific company, that's what our [research reports](/research) deliver.

---

*This is an investing framework for educational purposes only. Not investment advice. The author may hold positions in securities discussed. Always do your own due diligence before making investment decisions.*


Most management scorecards are one-size-fits-all. Mine changes entirely based on the type of business. Here's the archetype-first framework with real examples.

How I Grade Management Teams (It Depends on the Business)


Nearly every stock has a price where it's an attractive buy. The challenge is finding the ones where that price is *today's* price.

In **small cap investing**, these opportunities show up more often than most people realize. Higher volatility, thinner analyst coverage, less institutional participation, structural forced selling from spin-offs and index rebalancing, minimum market cap and price thresholds that lock out large funds — all of it creates mispricings that simply don't exist in large caps. A $50 billion fund can't take a meaningful position in a $200 million company. That structural gap is where I spend most of my time.

I look at over 200 ideas per week through Pelican's screening pipeline. Before any of them get serious research time, they have to pass three questions. But the way I apply those questions depends on *what kind* of opportunity I'm looking at. I invest in two distinct buckets — **GARP** and **deep value** — and while the three questions are the same, the answers look very different.

---

### The Two Buckets

**GARP (Growth at a Reasonable Price)** is where I look for high-quality businesses trading at reasonable multiples. These are companies with high ROIC, durable competitive moats, and secular tailwinds — but something temporary has compressed the valuation. The bet is that the market is overweighting a short-term problem and underweighting the long-term earning power.

**Deep value** is where I look for stocks trading below what the assets are worth today. Low price-to-tangible-book, heavy insider ownership, commodity-tied cyclical businesses where the market is pricing in permanently depressed conditions. The bet here isn't on growth — it's on reversion to the mean and the margin of safety embedded in hard assets.

Both strategies share the same core belief: the market misprices small caps more frequently and more severely than large caps. But the toolkit for each is different. Let me walk through both.

---

### Question 1: Is the Valuation Actually Attractive?

This is the entry point for everything I do. If the price isn't compelling relative to what I'm buying, nothing else matters.

**For GARP**, I'm looking for quality businesses trading under 20x EBITDA with strong returns on invested capital. High ROIC is a proxy for business quality — it tells you the company is generating real economic value, not just revenue. I want to see historical revenue growth, secular tailwinds in the industry, and a **competitive moat** — network effects, switching costs, scale advantages — that protects margins over time.

**For deep value**, the bar is different. I want to see a **price-to-tangible-book near or below 1x**, the stock trading below peer multiples, and ideally hard assets that the market is undervaluing — long-owned real estate, idle plants, fleet assets, or inventory that could be liquidated for more than the market cap implies.

In both cases, I'm also looking for **high insider ownership**. I want management teams whose net worth rises and falls with the stock price — not hired-gun CEOs collecting a salary regardless of performance. Aligned incentives change how capital gets allocated.

#### What kills an idea at Question 1

**Secular decline.** A cheap stock in a structurally shrinking business is almost never a good investment, regardless of the multiple. Take BTMD — a hormone optimization platform trading at roughly 2x earnings with Q4 2025 procedure revenue declining 13% as clinic attrition accelerates. The P/E screams "cheap," but the core business is contracting. No amount of valuation discount compensates for a melting ice cube.

---

### Question 2: Why Does This Opportunity Exist?

If a stock looks cheap, there's a reason. You need to understand that reason before you can determine whether it's a real opportunity or a trap.

I'm looking for a **legible reason for mispricing** — something specific and identifiable that explains the discount, and that I believe the market has wrong. The best small cap opportunities come from a handful of recurring patterns:

- **Neglect.** No analyst coverage, too small for institutional mandates, recently spun off with no natural shareholder base. The stock isn't mispriced because someone analyzed it and disagreed — it's mispriced because nobody's looked.
- **Temporary operational drag.** A bad quarter, a money-losing division, a one-time charge that the market extrapolates as permanent.
- **Cyclical trough.** A commodity-tied business at the bottom of its cycle, where the market is pricing in permanently depressed margins.
- **Sector contagion.** The entire sector sold off on a macro headline, and the strongest operators got unfairly punished.
- **Complexity.** A restructuring, divestiture, or management transition that makes the financials messy and hard to model. Most analysts give up. The few who do the work find the value.

If I can't articulate a specific, falsifiable reason why the market is wrong, I don't have a thesis — I just have a cheap stock.

#### GARP Example: ThredUp (TDUP)

**TDUP** | Online Resale / Consumer Marketplace | Temporary operational drag

![ThredUp (TDUP) stock chart — $0.53 low in November 2024 to $12.08 high in August 2025](/insights/charts/tdup-chart.png)

ThredUp is one of the largest online resale platforms in the US and the largest *managed* resale marketplace — sellers ship in clothing and ThredUp handles pricing, photography, listing, and fulfillment. The US-only business runs at **~79% gross margins** with 20% revenue growth and strong secular tailwinds (the global secondhand apparel market is projected to reach $486B by 2031 per Research and Markets, growing at ~16% annually).

So why was the stock trading at $0.53 in November 2024 — down 98% from its IPO high?

The answer was **Remix**, a European resale platform ThredUp acquired for $28.5M in 2021. Remix ran at 24% gross margins versus 80% in the US and was bleeding cash at a negative 23% adjusted EBITDA margin. The European division was masking the quality of the core business. Every consolidated financial statement made ThredUp look like a money-losing operation with no path to profitability.

**The mispricing was legible:** the market was valuing the whole company based on blended metrics that included a division destroying value. Strip out Europe, and you had a high-margin marketplace with network effects, 163 brand partners on its Resale-as-a-Service platform, and a durable competitive moat built on $400M+ in proprietary reverse logistics infrastructure.

**Before exit (Nov 2024)** → **After exit (2025):**

- **Stock price:** $0.53 (all-time low, Nov 4) → $12.08 (52-week high, Aug 25)
- **Gross margin (US):** 80% (masked by Europe's 24%) → 79.4% (reported clean)
- **Adj. EBITDA margin:** Negative (blended) → 4.4% and expanding
- **Active buyers:** Declining → 1.65M (+30% YoY, record)
- **Annual cash flow:** Negative → +$3.1M (first-ever positive)

ThredUp announced the European exit in August 2024 and completed the divestiture in early December. The stock bottomed at $0.53 on November 4, started 2025 at $1.41, and ran to $12.08 by late August 2025 — roughly **23x from the absolute low** and **8.5x from the start of 2025**. First positive cash flow year in company history. The business didn't change. The market's ability to see it clearly did.

---

### Question 3: What Re-Rates This Stock?

A stock can be cheap for years. A business can be improving for years. Neither fact alone puts money in your pocket. Something has to change the market's perception — and ideally, I can identify what that something is before it happens.

This is the **catalyst question**, and it's where I separate ideas I watch from ideas I put capital behind.

Good catalysts are specific and time-bound:

- **Earnings inflection.** Consensus estimates haven't caught up to improving fundamentals. This is the most common catalyst in small caps because analyst coverage is thin and estimate revisions lag.
- **Insider buying clusters.** Management and board members buying stock with their own money — not exercising options, writing checks. As I covered in [How to Read 13F Filings](/blog/how-to-read-13f-filings), following smart money works when you understand the context. Insider buying in a small cap with improving fundamentals is one of the strongest confirming signals I know.
- **Activist involvement.** An activist investor taking a position and pressuring for change — asset sales, board seats, strategic review. When I was at Osmium Partners, we ran these campaigns ourselves. As a passive investor now, I look for situations where someone else is creating the catalyst.
- **Capital return.** Buybacks or dividend initiation signal management confidence and attract new investors.
- **Strategic activity.** M&A in the sector, a competitor getting taken private, or comparable transactions that highlight the valuation gap.

The key word is **specific**. "The market will eventually realize this is cheap" is not a catalyst. "Crush margins just hit $0.91/gallon — up 435% from last year — and Q3 earnings will reflect it before the street updates estimates" — that's a catalyst with a date and a mechanism.

#### Deep Value Example: Alto Ingredients (ALTO)

**ALTO** | Ethanol / Renewable Fuels | Cyclical trough + activist catalyst

![Alto Ingredients (ALTO) stock chart — $0.76 low in April 2025, recovering to $4.30+ by March 2026](/insights/charts/alto-chart.png)

ALTO is an ethanol producer — a commodity-tied, cyclical business with **~350 million gallons of annual production capacity** across five plants. This isn't a new pattern for ALTO — in the COVID crash of 2020, the stock traded down to a $14M market cap before rallying over 3,500% as crush margins recovered. The market has a habit of pricing ethanol producers for death during cyclical troughs, then scrambling to reprice them when margins revert.

By late 2024, crush margins had compressed again and ALTO was trading near $1.68 — well below tangible book value. Activist investor **Kaufman Kapital** (~4.9% ownership, the largest stockholder) published a letter in December 2024 estimating ALTO's plants were collectively worth **$550M+ (~$6/share)** and pressured the board to explore strategic alternatives. For context, when GEVO announced the acquisition of Red Trail Energy's 65M gallon ethanol plant in September 2024 for $210M (including carbon capture assets, deal closed February 2025), the implied valuation was **$3.23/gallon of capacity**. Even discounting for the CCS premium, that comp implies ALTO's capacity is worth multiples of its market cap.

**Late 2024 (Kaufman letter)** → **March 2026:**

- **Stock price:** ~$1.68 → ~$4.30
- **Kaufman NAV estimate:** $6/share (still implies upside)
- **Price vs. Kaufman NAV:** ~0.28x of estimated asset value → ~0.72x
- **Q4 net income:** -$42.0M (Q4 2024) → +$21.5M (Q4 2025)
- **Crush margins:** Compressed → Recovering

What was the catalyst? **Activist pressure + earnings inflection.** Board changes followed Kaufman's letter — new chair and vice-chair installed in June 2025. Then crush margins started recovering, and Q4 2025 delivered $21.5M in net income versus a $42M loss in the prior year's quarter. The stock bottomed at $0.76 in April 2025 and has since rallied to $4.30 — a **5.7x move** in under a year, with the activist's $6/share NAV estimate still implying meaningful upside.

This is what deep value looks like in practice: hard assets trading at a fraction of replacement cost, an activist creating urgency, and a cyclical turn delivering the earnings catalyst the market needed to re-rate.

---

### Deep Value Requires More Nuance

I want to be honest about this: deep value is harder than GARP. There are more ways to lose money. The checklist is longer and the margin for error is thinner.

Beyond the three questions, here's what I stress-test in every deep value situation:

- **Zero solvency risk.** I sensitize commodity prices to worst-case scenarios. What would it take for this company to go bankrupt? If the answer involves even moderately adverse conditions — negative free cash flow, heavy debt maturities, rising interest rates — I pass. The whole thesis depends on surviving to see the reversion.
- **Not in secular decline.** This is the single most important filter. Cheap cyclical businesses can recover. Cheap businesses in structural decline usually don't. I need to believe the industry has a future, not just a past.
- **Hidden or undervalued assets.** Long-owned real estate carried at historical cost, fleet assets worth more as scrap than the market cap implies, idle capacity that has option value if conditions improve.

#### Deep Value Example: Imperial Petroleum (IMPP)

**IMPP** | Shipping (Tankers + Drybulk) | Asset discount + governance overhang

![Imperial Petroleum (IMPP) — trading at ~59% discount to NAV as of March 2026](/insights/charts/impp-chart.png)

IMPP is a Marshall Islands-incorporated shipping company headquartered in Athens that illustrates both the opportunity and the complexity of deep value. Here are the numbers as of March 2026:

- **Stock price:** ~$4.64 | **Market cap:** ~$206M
- **Management NAV estimate (Sept 2025):** $508M ($11.38/share basic, $9.21 fully diluted) — a **~59% discount** on basic shares
- **Cash + time deposits:** $179M (year-end 2025) | **Total debt:** $0
- **Fleet:** 20 vessels (expanding to 26) | **Vessels book value:** ~$343M (Q3 2025)
- **P/Tangible Book:** ~0.33x | **EV/EBITDA:** ~0.38x
- **FY2025 net income:** $50M | **Insider ownership (Vafias family):** 30.1%

Read those numbers again. You can sell the ships for scrap, add the cash balance, and the sum exceeds the stock's market cap. The enterprise value is ~$27M for a fleet generating $50M in annual net income. At 0.33x tangible book, you're buying dollars for thirty-three cents.

**Why does this opportunity exist?** Governance risk and serial dilution. The Vafias family controls 30% of the stock, all management services flow through family entities, and the share count has more than doubled since a 2023 reverse split — including a recent $60M offering with warrants that could add another ~19M shares (~43% dilution from current levels). The market is pricing in continued dilution and related-party extraction. It's a legitimate concern.

**What could re-rate it?** The company initiated a new $10M buyback program in February 2026. If management follows through — buying back shares at 0.33x book instead of diluting — it would signal a shift in capital allocation philosophy. Meanwhile, the fleet is expanding to 26 vessels with zero debt, and tanker/drybulk rates remain supportive.

This is the tension in deep value: the assets are real, the discount is enormous, but the catalyst depends on management behavior that contradicts their track record. It requires more nuance, more monitoring, and a clear-eyed view of the risks. Not every deep value stock works out. But when the margin of safety is this wide, you don't need everything to go right — you just need the worst case not to happen.

---

### The Filter in Practice

These three questions work as a sequential funnel, applied differently depending on the bucket:

**Q1 — Valuation:**
- **GARP:** Under 20x EBITDA, high ROIC, secular tailwinds
- **Deep Value:** Below tangible book, hard asset coverage, peer discount

**Q2 — Why mispriced?**
- **GARP:** Temporary drag masking quality (TDUP's Europe division)
- **Deep Value:** Cyclical trough, neglect, forced selling

**Q3 — Catalyst:**
- **GARP:** Earnings inflection, operational cleanup
- **Deep Value:** Activist pressure, commodity reversion, asset sales

**Kill switch:**
- **GARP:** Secular decline, no moat
- **Deep Value:** Solvency risk, secular decline, no asset floor

Most ideas die at Question 1 — the valuation either isn't compelling or the business is in secular decline. Of the survivors, maybe half have a legible mispricing reason. Of those, maybe half have a real catalyst. A list of 200 weekly ideas typically produces 3-5 names worth deep underwriting.

That's by design. In small cap investing, the edge isn't seeing more ideas — it's killing bad ones faster. The three-question filter is how I protect my scarcest resource: the time I spend on research that actually leads to conviction.

If you want to see this framework applied to real companies — with full financial analysis, variant perception scoring, and catalyst identification — that's what [Pelican's research reports](/research) are built to deliver.

---

*This is an investing framework for educational purposes only. Not investment advice. The author may hold positions in securities discussed. Always do your own due diligence before making investment decisions.*


I invest in small caps two ways — GARP and deep value. Here are the 3 questions I use to separate real opportunities from value traps, with real case studies.

3 Questions I Ask Before Every Small Cap Investment


Block cut roughly 4,000 employees yesterday — nearly half its entire workforce. The stock jumped 24% after hours.

Jack Dorsey didn't frame this as a restructuring. He didn't call it a "realignment." He said AI can do the work, and he expects the majority of companies to reach the same conclusion within the next year.

And I don't think Dorsey is speculating from the sidelines. He's a founder-operator deeply embedded in Silicon Valley. When he says he expects others to follow, I am assuming he's had real conversations about this with other founders and CEOs. This isn't a prediction — it reads more like a preview.

The market's reaction tells you everything you need to know. Investors didn't see a company in distress. They saw a company that just permanently expanded its margins.

---

### This Isn't a Normal Layoff Cycle

The 2022-2023 layoff wave was about unwinding pandemic over-hiring. Demand slowed, companies right-sized. That's cyclical.

This is structural. Dorsey isn't saying demand is down — he's saying AI replaces the work itself. That distinction matters enormously for investors. Cyclical cost cuts are temporary. Structural margin expansion reprices a company permanently.

And the timing matters. We're past the "AI is coming" phase. We're in the "AI is here and it's cheaper than your employees" phase. Block just proved the playbook: cut aggressively, invest in AI, and the market hands you a 24% premium overnight.

---

### The Playbook: Three Tests

The companies best positioned to benefit from AI-driven labor reduction need to pass three tests:

**Test 1: High labor costs as a percentage of revenue.** This is table stakes. If a company's workforce isn't a material cost driver, cutting headcount doesn't move the needle. You want businesses where labor is 30%, 40%, 50%+ of the cost structure — that's where the margin expansion math gets real.

**Test 2: Someone with the power and incentive to force the decision.** This is the most underappreciated part of the thesis. Dorsey is a founder-operator. His net worth moves with Block's stock price. That alignment gave him the conviction to make a move this bold — cutting half your workforce makes headlines, and not the comfortable kind.

Compare that to a career CEO earning a flat salary with modest equity. They have every incentive to play it safe, avoid the headline risk, and keep headcount stable. The companies that move first on AI-driven labor reduction will be the ones where *someone* has both the power and the financial incentive to pull the trigger. That could be a founder-operator like Dorsey or Zuckerberg. It could be an activist investor — an Elliott Management or a Starboard Value — who can credibly walk into a board meeting and say "your competitor just cut 40% of headcount and the stock jumped 24%, why haven't you?" I spent time at an activist fund earlier in my career, and I can tell you — this is exactly the kind of data point that gets turned into a slide deck and presented to a board within a week. It could be a controlling shareholder or a concentrated ownership structure. But someone needs to have skin in the game and the ability to force the hard call. Without that catalyst, most boards will default to inertia.

**Test 3: A durable competitive moat that AI enhances rather than erodes.** This is where most people stop thinking too early. Not every company benefits from AI — some get disrupted by it.

We saw this play out when Anthropic launched Claude with autonomous coding capabilities. Tech stocks sold off because the market started pricing in a future where companies need fewer software engineers — which means they also need fewer developer tools, fewer cloud compute instances, and fewer seats of every SaaS product engineers use daily. The companies that were supposed to benefit from AI were suddenly threatened by it.

That same logic applies broadly. If an IT outsourcing firm's clients can use AI agents to do the work they currently pay the outsourcer to do, that's bearish — even if the outsourcer's own internal margins improve. If AI video tools get good enough, that doesn't just help a studio cut costs — it enables competitors to produce content at a fraction of the cost, compressing the incumbent's moat.

The key question for each name isn't just *"can this company use AI to cut labor costs?"* It's *"does AI also erode their competitive advantage?"* Companies with durable moats — network effects, regulatory capture, physical infrastructure, proprietary data — benefit cleanly. Companies whose moat is primarily *"we have a lot of smart, expensive people"* are more exposed than they appear. AI helps them in the short term (fewer people, same output) but threatens them in the long term (their clients realize they don't need them at all).

---

### Applying the Framework

Here's how I think about applying this across a few categories:

**Where all three tests pass cleanly.** Companies like Amazon (AMZN) and Meta (META) come to mind — massive workforces, founder-operators or equity-aligned leadership, and moats built on physical infrastructure or network effects that AI strengthens rather than undermines. These are the types of businesses where labor reduction flows straight to margins without opening the door for competitors.

**Where activist pressure could be the catalyst.** I think this is an underappreciated angle. Companies under activist pressure already have external forces demanding cost cuts. AI gives those activists a new and powerful quantifiable argument. Look for names where an activist holds a significant stake, the stock has underperformed, and the underlying business has a real moat (marketplace network effects, proprietary data, etc.) that AI enhances. Etsy is an interesting example here — Elliott Management holds ~13% as the largest shareholder, they already forced a CEO change effective January 2026, the stock is down 60% in three years, and the marketplace of unique sellers is a genuine network effect that AI could help serve more efficiently. The external pressure plus the underlying moat plus the room to compress operations is a combination worth watching.

**Where it looks like a beneficiary but isn't.** IT outsourcing firms are the clearest example. Hundreds of thousands of employees doing code migration, back-office processing, consulting deliverables — exactly the work AI agents are getting good at. On the surface, it looks like they pass test one easily. But they fail test three entirely. Their moat *is* their people, and their clients can increasingly replace them with AI tools directly. That's not a labor reduction opportunity — it's an existential threat disguised as one.

---

### Second-Order Thinking: Who Gets Hurt

This is where it gets interesting if you're looking at the other side of the trade.

**IT Services & Outsourcing** — The firms whose moat is fundamentally *"we have a lot of smart people."* AI could improve their internal margins in the short term, but the real risk is their clients using AI agents to cut them out entirely. This category looks like a beneficiary on the surface but fails the third test.

**Commercial Real Estate** — Fewer workers means less office space, period. The sector was already under pressure from remote work. AI-driven headcount reductions accelerate the secular decline. This isn't a temporary vacancy problem — it's permanent demand destruction.

**Staffing Firms** — Their entire business model is placing humans in jobs. If AI structurally reduces the number of knowledge-work positions, these companies face an existential headwind. The market hasn't fully priced this in.

**Per-Seat SaaS** — If headcounts shrink 30-40% industry-wide, every company charging per-seat pricing sees revenue compress proportionally. Unless these companies pivot to usage-based or outcome-based pricing, their revenue base is structurally impaired.

---

### What I'm Actually Watching

**Margins, not announcements.** The proof is in whether companies that cut headcount actually flow those savings to the bottom line — or just reinvest them into AI infrastructure spending. Block's next two quarters will be the most important data point here. If operating margins expand 500+ bps with half the headcount, every board in America takes notice. That's the catalyst that turns one company's decision into an industry-wide movement.

**Who's running the company.** When I'm evaluating which companies might follow Block's lead, I look at who actually has the power and incentive to make the call. Founder-operators with real skin in the game are the ones who will move first. Activist investors with significant stakes will push next. A career CEO with a $2M base salary and a golden parachute has no incentive to risk the PR firestorm. Follow the incentives.

**The market's reaction to the next one.** Block got 24%. If the next company to announce a 30%+ AI-driven workforce reduction gets the same reaction, this becomes a full-blown playbook that spreads fast. If the market shrugs, the thesis needs recalibrating. My bet is the market rewards it again. And again.

**Per-seat SaaS pricing model shifts.** This one is more speculative, but I think it could be a useful leading indicator. If a major SaaS company announces a move to usage-based pricing, that's the market quietly admitting that headcount compression is real and accelerating. It hasn't happened yet at scale, but it's the kind of structural signal I'd want to catch early.

---

### The Bottom Line

Block's announcement isn't the story. It's the starting gun.

Dorsey said it publicly, but this conversation has been happening privately in board rooms for months. He's too connected and too deliberate for this to be an isolated decision made in a vacuum. The companies that move first will get rewarded by the market — we literally just watched it happen in real time, 24% in a single session. The ones that wait will eventually get forced into it by competitive pressure, by activist investors, or by AI-native competitors who never had the bloated headcount to begin with.

The framework is straightforward: look for companies with high labor intensity, durable competitive moats that AI *enhances* rather than erodes, and someone in the room — a founder, an activist, a controlling shareholder — with the incentive and the power to actually make the hard call. Avoid the companies on the wrong side — the staffing firms, the per-seat SaaS names, the outsourcing giants, and anyone whose moat is "we hire smart people."

The labor market is about to go through a structural repricing. Position accordingly.

---

*Disclaimer: This is market commentary for educational purposes only. Not investment advice. I may hold positions in securities mentioned. Always do your own research.*


Block cut nearly half its workforce. The stock jumped 24%. Jack Dorsey says everyone else will follow. Here's who benefits, who gets hurt, and what it means for your portfolio.

Block Just Gave Every CEO in America Permission to Cut Their Workforce in Half


![How to Read 13F Filings — a framework for extracting real signal from institutional filings](/blog/13f-hero.png)

Every quarter, roughly 45 days after period-end, thousands of institutional investors file a document with the SEC called a 13F. It's a snapshot of their long equity holdings — what they own, how many shares, and the market value.

And every quarter, the financial media reduces it to the same headline: **"Buffett bought X. Ackman sold Y."**

That's not analysis. That's gossip.

I've spent the better part of a year building a system that tracks 13F filings from 45 of the best fundamental investors in the world — funds like Baupost, Pershing Square, Lone Pine, Appaloosa, Himalaya Capital, and Oaktree. Not to chase their picks, but to understand where conviction is clustering and capital is flowing.

Here's what I've learned about how to actually read these filings.

---

### What a 13F Filing Actually Is

A 13F-HR is a quarterly SEC filing required of institutional investment managers with at least $100 million in qualifying assets. It reports long positions in US-listed equities and certain other securities (options, convertible notes, etc.).

A few things to understand upfront:

- **It's backward-looking.** A filing for Q4 2025 (positions as of December 31) isn't due until February 14, 2026. By the time you're reading it, the fund may have already changed its position. This 45-day lag is the single most important thing to internalize.
- **It only shows long equity positions.** You won't see short positions, fixed income, private holdings, derivatives (mostly), or cash. You're seeing a partial picture.
- **It reports market value, not cost basis.** You can see what a position is worth, but not what they paid for it or when they started buying.

None of this makes 13Fs useless. It just means you need to know what to look for.

---

### Mistake #1: Focusing on Individual Buys

This is the most common error. A headline says "Baupost opened a new position in Amazon" and people rush to buy AMZN. But context matters enormously.

In our Q4 2025 analysis, Baupost's new Amazon position was roughly $490K — representing just 9.3% of their portfolio. For a fund of Baupost's size, that's a meaningful allocation. But if it had been 0.1% of their portfolio? That's a rounding error. Maybe it's a hedge, a basket trade, or an analyst's exploration position.

**What actually matters:** the position size relative to the fund's total AUM (assets under management) and the number of holdings they run. A $50 million fund putting 15% into a single stock is making a statement. A $20 billion fund putting 0.03% into that same stock is barely whispering.

---

### What to Actually Look For

After processing hundreds of 13F filings, I've found five patterns that consistently produce actionable signal. These aren't theoretical — they're the exact scoring framework I use.

#### 1. Buying Clusters

A single fund opening a new position tells you something. Three or four unrelated funds independently opening new positions in the same stock during the same quarter tells you something much louder.

These are **buying clusters** — multiple funds arriving at the same idea independently. They didn't coordinate (that would be illegal). They each did their own work and reached the same conclusion at roughly the same time.

In Q4 2025, Amazon scored a 9 out of 10 on our system. Not because Baupost bought it — but because three separate funds initiated new positions (Baupost, Gardner Russo & Quinn, and Wedgewood Partners) while four more added to existing positions (Pershing Square increased by 65%, Peconic Partners by over 7,500%, Kensico by 5%, and NWI Management by 176%).

Seven funds. All moving the same direction. That's a cluster.

![AMZN Q4 2025 — 7 funds active, 3 new buys, 4 additions, $968M+ total capital deployed](/blog/13f-amzn-cluster.png)

#### 2. High-Conviction Weighting

When a fund makes a stock one of its largest positions — say, over 10% of its total portfolio — it's expressing serious conviction. These managers live and die by their performance. They're not putting 15-20% of their fund into something they're lukewarm about.

In the same Q4 report, Himalaya Capital held Alphabet (GOOGL) at 21.6% of their portfolio. Pershing Square had it at 12.5%. Ruane Cunniff (the Sequoia Fund team) and Abrams Capital both held it at 10.6%.

Four different high-caliber funds, all with 10%+ weights in the same stock. That's convergence.

#### 3. The "Mismatched Size" Signal

This is my favorite and the one that most 13F analysis completely misses.

When a large fund ($5-20 billion AUM) or a mega fund ($20+ billion) takes a position in a small-cap ($500M-$2B) or micro-cap (under $500M) company, pay attention. These funds have analysts, compliance teams, and investment committees. They don't accidentally buy a $300 million market cap company.

The signal is even stronger when the fund's AUM dwarfs the company's market cap. A $7 billion fund buying into a $292 million company (like Donald Smith's new position in James River Group Holdings) means that fund had to actively choose to fish in a very small pond. The due diligence bar for that decision is extremely high.

Our scoring system gives extra points for these mismatched-size signals — and adds a kicker for companies under $100 million in market cap, where the asymmetry is most extreme.

#### 4. Position Increases (Not Just New Buys)

New positions get all the attention. But a fund that already owns a stock and increases its position by 40-65% is often a stronger signal than a brand-new purchase.

Why? Because they already know the company. They've done the work. They've sat through earnings calls, talked to management, and watched the thesis play out. And after all that, they want *more*.

In Q4 2025, Brookfield Corp (BN) was a great example. Akre Capital increased its position by 40% (to 13.5% of its portfolio). Lone Pine added 56%. Third Point added 33%. Pershing Square added 50%, bringing its position to $935 million — 18.1% of its entire fund. Four separate funds increasing the same position in the same quarter.

#### 5. "Conviction Holds"

Not every signal involves buying. Sometimes the most telling data point is what funds *didn't* sell.

When a stock declines and a fund holds its position unchanged — especially when that position is 10%+ of their portfolio — that's information. They had every reason and opportunity to trim, and they chose not to. In a world where most portfolio managers are incentivized to cut losers and window-dress their reports, the decision to hold through drawdowns reveals conviction.

Our Q4 report flagged five "conviction holds" — positions where elite funds maintained large allocations despite meaningful price declines since the prior quarter's filing.

---

### The 45-Day Lag: Feature, Not Bug

Everyone complains about the delay. "By the time you read the 13F, the trade is old news." This is true — and it doesn't matter as much as you think.

Here's why: the investors I track aren't day traders. Baupost, Pershing Square, and Himalaya Capital hold positions for years. A new position initiated in Q4 2025 isn't a trade that'll be gone by Q1 2026. It's the beginning of a thesis that often takes 12-24 months to play out.

The 45-day lag matters for momentum traders. It barely matters for fundamental investors following other fundamental investors.

What I actually do with the lag: I compare the stock's price at quarter-end (when the filing reflects their holdings) to the current price. If a fund opened a new position at $20 and the stock is now at $17, the setup might be *better* than when they bought it. The thesis doesn't change — but you're getting a lower entry point than the fund manager did.

In our Q4 2025 report, Amazon had drifted -11.1% from its filing-date price. FICO was down -19.1%. S&P Global was down -19.7%. If the thesis that attracted multiple elite funds was valid at higher prices, a double-digit decline doesn't invalidate it — it amplifies the opportunity.

---

### How I Built This Into a System

Reading one 13F filing is straightforward. Reading 45 of them every quarter, cross-referencing the holdings, identifying clusters, weighting by conviction, and flagging size mismatches — that's a different problem entirely.

I built a pipeline that does this programmatically:

1. **Download and parse** every 13F-HR filing for 45 tracked funds directly from SEC EDGAR
2. **Aggregate by security** — for each stock, map out every fund that holds it, every new buy, every addition, and every conviction hold
3. **Score each position** across five dimensions: holding convergence, buying convergence, high-weighting conviction, position increases, and mismatched-size signals
4. **Rank by total score** on a 1-10 scale and generate a signal report with the full detail behind each rating

![The 5-dimension scoring framework Pelican uses to rank 13F signals](/blog/13f-scoring-framework.png)

The result is a quarterly report like our [Q4 2025 13F Institutional Flows analysis](/13f-latest) — not a list of "what Buffett bought," but a systematic map of where the smartest money is clustering.

---

### The Bottom Line

13F filings are one of the few free, reliable sources of institutional investment data available to individual investors. But most people extract maybe 5% of the available signal because they stop at the headline.

The real value isn't in any single fund's single purchase. It's in the patterns across funds: where conviction is clustering, where large funds are entering small companies, where existing holders are doubling down, and where everyone is holding through volatility.

That's what I track. That's what our [13F signal reports](/13f-latest) are built to surface.

If you want to dig into the raw data yourself, start with the SEC's [EDGAR full-text search](https://efts.sec.gov/LATEST/search-index?q=%2213F-HR%22&dateRange=custom&startdt=2026-01-01&enddt=2026-02-28) for 13F-HR filings. Or check our [methodology page](/methodology) for a deeper look at how the scoring system works.

Either way — read the actual filings. The signal is there. You just have to know where to look.

---

*This analysis is for educational purposes only and does not constitute investment advice. The author may hold positions in securities discussed. Always do your own due diligence before making investment decisions.*


Most 13F analysis stops at 'Buffett bought Apple.' Here's how to actually extract signal from institutional filings — using the same framework I built to track 45 of the world's best investors.

Blog

Why I Built Pelican

How I Grade Management Teams (It Depends on the Business)

3 Questions I Ask Before Every Small Cap Investment

Block Just Gave Every CEO in America Permission to Cut Their Workforce in Half

How to Read 13F Filings (And What They Actually Tell You)

Get Daily Signals in Your Inbox