Cluster Buying: When Multiple Insiders Buy Together

What Cluster Buying Is

One insider buying stock tells you something. Three insiders buying in the same month tell you considerably more. That difference is the whole reason this page exists.

Cluster buying is the pattern where two or more insiders at the same company make open-market purchases within a short window, usually 30 days. Each of those people has a different vantage point on the business: the CFO sees the numbers, division heads see operations, directors see the strategy debates. When several of them reach for their own checkbooks in the same month, each has independently concluded the stock is worth more than it costs.

Why It Beats a Single Purchase

A lone purchase is easy to explain away. Maybe the CEO is signaling confidence for optics. Maybe a director is topping up to meet an ownership requirement. Maybe someone simply had cash on hand. Any single trade carries that kind of noise.

Clusters mostly don't. The odds of three people independently deciding to buy for unrelated personal reasons in the same month are slim, and this has been documented for a long time. Jaffe (1974) showed that insider trading's predictive power strengthens markedly when several insiders at the same firm trade together, and later research has kept confirming it.

That said, "stronger" is not "infallible." We've watched plenty of clusters fizzle. Most of this guide is about telling the difference.

How We Detect Clusters on InsiderAction.io

The live snapshot above is not a static list. It is recomputed from raw SEC Form 4 filings every hour. Our detection rule is deliberately strict so the signal stays clean:

• Two or more distinct insiders (matched by name and CIK, so one person filing twice never counts as a cluster)
• Open-market purchases only: transaction code P on non-derivative securities. We exclude option exercises, awards, and gifts, which inflate raw "insider buying" counts elsewhere but carry no conviction signal.
• Inside a rolling 30-day window, keyed on the transaction date rather than the (often-delayed) filing date.

For every cluster we also compute two things most trackers don't: a dollar-weighted average ownership change, so a token $5,000 purchase doesn't get equal billing with a CEO backing up the truck, and the average ISS score of the participating insiders, our measure of how well each person's past trades have preceded price moves. A cluster of high-ISS buyers is a materially different signal from a cluster of insiders with poor track records, even when the headline "3 insiders bought" looks identical.

You can see the full, sortable list any time on the Clusters page.

Judging a Cluster

Two questions matter more than everything else: who is buying, and how much of their own money is at stake.

On the who: a cluster of C-suite executives outranks a cluster of directors, which outranks a cluster of lower-level insiders. Operating executives live inside the numbers; directors see them quarterly. A mix of both is a good sign. It means the conviction spans different vantage points on the business.

On the how much: dollar-weight everything. Three insiders committing a combined $2 million is a serious statement. Seven insiders committing $15,000 is paperwork (more on that below). Size the purchases against what these people presumably earn. A CEO putting a year of salary into the stock is telling you something a token buy never will.

Then add context. Buying into a decline, at 52-week lows, or during a broad selloff carries more information than buying after a run-up, because insiders stepping in front of a falling stock are taking reputational risk along with financial risk. And check the company's own history on its ticker page: if past clusters there preceded gains, and the current buyers have strong insider scores, the odds improve again.

Red Flags

The headline number ("5 insiders bought!") hides a lot of junk. Patterns we've learned to discount:

Token amounts. A snapshot of our own clusters page in the summer of 2026 had a small water utility near the top of the list with seven different insiders buying. Combined value: about $11,000. That isn't conviction. That's directors meeting minimum stock-ownership requirements. It's also why we show dollar values and ownership change next to every cluster instead of just the buyer count.

Same-day, same-size purchases. When every buy lands on one day for similar amounts, you're often looking at a company-organized program rather than independent decisions.

Buys right after grants. Purchases that closely follow equity awards may be about maintaining a required position, not a view on the stock.

Clockwork patterns. Trades that recur on a fixed schedule are usually 10b5-1 plans executing automatically, whatever the price.

A Real Example: UnitedHealth, May 2025

In May 2025, UnitedHealth Group (UNH) was in freefall. The stock had roughly halved from its late-2024 highs after the company pulled its annual guidance and its CEO abruptly departed. It was the kind of headline sequence that usually has everyone heading for the exits. Then, over three days, our transaction data for UNH recorded this:

• May 14: Directors John Noseworthy and Timothy Flynn bought a combined $580,000
• May 15: Director Kristen Gil bought $1.0 million at $271, near the bottom of the slide
• May 16: CFO John Rex bought $5.0 million. And Stephen Hemsley, who had returned as CEO only days earlier, bought $25 million at $288.57

Five insiders. $31.6 million. Three days. It checks every box discussed above: C-suite participation, serious money relative to any compensation package, independent decisions landing in a tight window, and buying straight into a collapse.

What happened next is instructive in both directions. Three months later, the stock was still down about 6%. Cluster buys don't call the bottom to the day, and anyone who bought expecting instant vindication was disappointed. But twelve months after the May 16 purchases, UNH was up roughly 36%, and the shares bought at the May 15 low were up more than 40%.

That's the honest version of what a strong cluster gives you: not a timing tool, but a well-supported claim that the people who know the company best believed the selling had gone too far. It took a year, not a quarter, to pay off.

Working Clusters Into Your Process

Clusters are a screening tool, not a buy list. The workflow we'd suggest: keep an eye on the Clusters page, or add tickers you follow to your watchlist so new filings come to you. When a cluster appears, spend ten minutes on the ticker page. Who bought? How much? Where in the price chart? With what track record? If it survives that look, do your normal fundamental homework on valuation, competitive position, and upcoming catalysts.

Put differently: a strong cluster earns a company your research time. It shouldn't spend your money by itself.

And when the buying spreads beyond one company, with several firms in the same industry clustering over consecutive months, it shows up on our Industry Streaks page. That is its own, slower-moving signal.