CLUSTER BUYING: A STEP-BY-STEP EXAMPLE

Reading about cluster buying in the abstract is one thing. Seeing it step by step is another. This guide walks through what a genuine cluster buying event looks like in SEC Form 4 data, how to evaluate the quality of the signal, and how to separate a real cluster from the look-alike patterns that produce false positives.

If you have not read the background piece, what is cluster buying and why does it matter covers the research and concept. This article is the applied version.

WHAT YOU ARE LOOKING FOR

A genuine cluster buy involves multiple distinct insiders at the same company making independent open market purchases within a compressed timeframe. The key word is independent. If the company ran a stock purchase plan that processed 15 employee purchases on the same day at the same price, that is not a cluster. It is a payroll event.

The factors that distinguish a genuine cluster from noise:

• Multiple distinct insiders with different roles and titles
• Independent timing: ideally different days, or at minimum different filing times
• Different prices: each buyer paid a different market price, meaning they chose to buy on different days
• No concurrent company-wide program that could explain the coordinated activity
• No concurrent selling from other insiders at the same company

A WORKED EXAMPLE

Here is a hypothetical but representative example of what a genuine cluster looks like in the data. The company is a mid-cap industrial manufacturer trading near a 52-week low after a sector selloff.

Let us score this across the key factors.

SCORING THE CLUSTER

This is a high-quality cluster. Four distinct insiders spanning the C-suite, the board, and operations all bought independently over eight days, at different prices, with no concurrent selling from other insiders. Each purchase represents a meaningful increase to their individual holdings. The total committed capital is nearly $1 million from people who all have better visibility into the company's near-term outlook than the market does.

WHAT A FALSE POSITIVE LOOKS LIKE

Here is a pattern that looks like a cluster but is not:

Every single purchase: the same date, the same price, the same share count. This is a company stock purchase plan or ESPP. HR processed it in batch on the 15th of the month. The insiders did not independently decide to buy stock at the same moment. This carries zero signal value regardless of how many names appear. InsiderTape's cluster detection engine applies a penalty to patterns like this and labels them accordingly.

HOW TO EVALUATE A CLUSTER ONCE YOU FIND ONE

Finding a genuine cluster is the beginning of the work, not the end. Before acting, run through these questions:

1. Does the company have obvious fundamental problems that insiders might be blind to? Insiders know their company's internal state well, but they may not fully appreciate competitive threats, secular headwinds, or macro risks that the market is pricing in correctly.
2. What was the stock doing before the cluster? Insiders buying after a 60% decline driven by a sector rotation are different from insiders buying into ongoing earnings disappointments. The former is often mean reversion; the latter is catching a falling knife.
3. Is there a catalyst on the horizon that could unlock value? A cluster in a company undergoing a strategic review, with an earnings call coming up, or with upcoming FDA approval is more timely than a cluster in a company with no obvious near-term catalyst.
4. What does the balance sheet look like? Insider buying is less meaningful if the company has a debt load that could impair equity even if the business recovers.

For a full framework on how to use insider data beyond just finding clusters, see our guide on how to actually use insider trading data.