We processed over 4 million SEC insider trading filings spanning over a decade. The dataset covers every Form 4 filing from 2015 through the present: every purchase, sale, grant, and exercise by directors, officers, and major shareholders of US public companies.

Most of what we found was noise, but the subset that represents genuine conviction has historically outperformed the market by a wide margin.

The noise breakdown

About 85% of all SEC insider filings are noise. Every insider transaction falls into a category based on its transaction code, and when you look at over 4 million filings in aggregate, the distribution is striking:

Sales (dispositions): ~40%

The largest category. Executives selling shares they already own. This sounds alarming (nearly 40% of all insider activity is selling) but most of it is routine. Executives receive a significant portion of their compensation in stock. They sell to diversify, to pay taxes, to fund life expenses. Many sales happen through pre-arranged automatic selling schedules set months or years in advance.

A CEO selling shares is not inherently bearish. Without knowing whether the sale was planned or discretionary, the filing alone tells you very little.

Stock grants and awards: ~35%

The second-largest category. These are shares awarded as part of compensation packages: restricted stock units (RSUs) vesting, performance share grants, director stock awards. They're not purchases. The executive didn't choose to buy. The company deposited shares according to a compensation agreement.

Grants and awards are pure noise from a signaling perspective. They tell you about a company's compensation structure, nothing about an insider's conviction.

Options exercises: ~10%

An executive exercises stock options that were granted as part of their compensation, converting them into shares. This is often immediately followed by a sale (exercise-and-sell) to cover the tax liability. Sometimes the executive holds the converted shares, but the exercise itself is a compensation event, not a conviction signal.

Most "insider buying alerts" that trigger on options exercises are misleading. The executive isn't putting new money in. They're converting an existing compensation instrument.

Open-market equity purchases: ~15%

That 15% is where the signal lives.

An open-market purchase means an insider went to their brokerage account, placed a buy order, and spent their own money to acquire shares at the current market price. No compensation plan triggered it. No automatic schedule executed it. The executive made a discretionary decision to increase their stake.

Only about 15% of all insider filings fall into this category. And it's within this 15% that the most studied, and most persistent, signals in financial research live.

Why most "insider buying alerts" are misleading

Finance media and social media treat insider filings as a monolith. A headline that says "Insiders Are Buying!" might be referring to a wave of options exercises, stock grants, or automatic plan purchases, none of which reflect an executive's personal conviction about their company's prospects.

The problem isn't that the data is wrong. The filings are accurate and timely. The problem is that the data is presented without context. A $5 million options exercise and a $5 million open-market purchase look the same in a raw filing feed. They are fundamentally different events.

The first is a compensation transaction that was likely scheduled months ago. The second is a CEO choosing to write a personal check at today's market price.

When we say 85% of insider activity is noise, this is what we mean. The raw feed includes everything, and most of it tells you nothing about what the executive thinks about their company's future.

What separates signal from noise

Within the 15% of filings that represent genuine open-market purchases, not all trades carry the same weight. Over a decade of data reveals several factors that distinguish the most informative signals:

Insider role matters. A CEO has the broadest view of a company's trajectory. A CFO knows the financial details intimately. A director may have strategic insight but less operational knowledge. Research consistently shows that purchases by C-suite executives, particularly CEOs and CFOs, carry more informational weight than purchases by directors or lower-level officers.

For a deeper look at how different insider roles produce different signal strength, see our article on which insiders produce the strongest signals.

Trade size relative to holdings matters. An executive who increases their holdings by 30% is making a different statement than one who increases by 2%. The absolute dollar amount matters too. A $2 million purchase by a CEO is a stronger signal than a $20,000 purchase by a director, but the holding increase percentage captures the personal conviction more precisely.

Cluster patterns matter. When multiple insiders at the same company buy within days of each other, the signal strengthens. This is called a "cluster buy." Three directors buying the same stock on the same day is not a coincidence. It suggests shared conviction based on shared information, the legal kind, the kind that comes from running a company.

Timing and market context matter. Insider buying during market downturns has historically been more informative than buying during rallies. When a CEO buys during a broad selloff, when everyone else is nervous, the signal is louder. For more on how timing affects insider signals, see our article on the timing edge.

Sector patterns

Insider buying is not evenly distributed across the market. Some sectors consistently generate more high-conviction signals than others, and the informational value of insider purchases varies by industry.

We covered this in detail in our article on insider buying by sector in 2026. The short version: sectors where executives have the clearest informational advantage, where they can see demand shifts, pipeline changes, or regulatory outcomes before the market, tend to produce the most informative insider signals.

The 15% that matters

The central finding from processing over 4 million SEC filings is simple but important: most insider activity is noise, and treating all filings equally is a mistake.

The raw data is public and free. SEC EDGAR publishes every Form 4 filing. Tools like OpenInsider display them in real time. The data has never been more accessible.

But accessibility without context creates a different problem. An investor checking insider filings sees hundreds of transactions per day (grants, exercises, automatic sales, compensation events) with no way to distinguish the routine from the meaningful.

That's what scoring solves. By evaluating every filing against over a decade of historical patterns (who bought, how much, in what context, and what happened next) it becomes possible to surface the small subset of filings that have historically been worth paying attention to.

Not every insider purchase is a buy signal. Not every insider sale is a sell signal. The data says something more nuanced: a specific type of insider purchase, by a specific type of executive, in a specific context, has historically been followed by returns that significantly exceed the broad market.

Finding those trades in a sea of 4 million filings is the problem. That's what InsiderSignals is built to solve.

See which insider filings our model scores highest. Start your free 30-day trial at insidersignals.io.

Related Reading

• How to Read a SEC Form 4 Filing: A step-by-step walkthrough of every field in the filing that US insiders must submit within 48 hours of a transaction.
• What Happens When You Invest $100 in Every High-Conviction Signal: The decade-long simulation showing cumulative returns for Elite and Strong tier trades versus the S&P 500.
• How We Score Every SEC Insider Filing: Inside the three-factor model that separates the 15% signal from the 85% noise.

For informational purposes only. Not investment advice.